OpenAI#Non-disparagement agreement

File:Pioneer Building, San Francisco (2019) -1.jpg in San Francisco]]

In December 2015, OpenAI was founded by Sam Altman, Elon Musk, Ilya Sutskever, Greg Brockman, Trevor Blackwell, Vicki Cheung, Andrej Karpathy, Durk Kingma, John Schulman, Pamela Vagata, and Wojciech Zaremba, with Sam Altman and Elon Musk as the co-chairs.{{Cite web |date=2024-07-12 |title=Only 4 of OpenAI’s 11 Founders Are Still With the Company—Where Are the Rest of Them? |url=https://observer.com/2024/07/openai-founders-career/ |access-date=2025-05-15 |website=Observer |language=en-US}} A total of $1 billion in capital was pledged by Sam Altman, Greg Brockman, Elon Musk, Reid Hoffman, Jessica Livingston, Peter Thiel, Amazon Web Services (AWS), Infosys, and YC Research.{{Cite web |date=December 12, 2015 |title=Introducing OpenAI |url=https://openai.com/blog/introducing-openai/ |access-date=December 23, 2022 |website=OpenAI |language=en |archive-date=August 8, 2017 |archive-url=https://web.archive.org/web/20170808104802/https://openai.com/blog/introducing-openai/ |url-status=live }}{{cite magazine |date=2015 |title=Sam Altman on His Plan to Keep A.I. Out of the Hands of the "Bad Guys" |url=https://www.vanityfair.com/news/2015/12/sam-altman-elon-musk-openai |url-status=live |archive-url=https://web.archive.org/web/20230203201103/https://www.vanityfair.com/news/2015/12/sam-altman-elon-musk-openai |archive-date=February 3, 2023 |access-date=January 23, 2023 |magazine=Vanity Fair}} The actual collected total amount of contributions was only $130 million until 2019. According to an investigation led by TechCrunch, while YC Research never contributed any funds, Open Philanthropy contributed $30 million and another $15 million in verifiable donations were traced back to Musk.{{cite web |title=Elon Musk used to say he put $100M in OpenAI, but now it's $50M: Here are the receipts |url=https://techcrunch.com/2023/05/17/elon-musk-used-to-say-he-put-100m-in-openai-but-now-its-50m-here-are-the-receipts/ |archive-url=https://archive.today/20230518211335/https://techcrunch.com/2023/05/17/elon-musk-used-to-say-he-put-100m-in-openai-but-now-its-50m-here-are-the-receipts/ |date=May 17, 2023 |archive-date=May 18, 2023 |work=TechCrunch |last=Harris |first=Mark}} OpenAI later stated that Musk's contributions totaled less than $45 million.{{Cite news |last=Chan |first=Kelvin |date=March 6, 2024 |title=OpenAI says Musk agreed the ChatGPT maker should become a for-profit company |url=https://apnews.com/article/openai-elon-musk-lawsuit-sam-altman-4a4c0a19316f849f65db9e6d2b0b7a6b |publisher=Associated Press}} The organization stated it would "freely collaborate" with other institutions and researchers by making its patents and research open to the public.{{cite web |title=Introducing OpenAI |url=https://blog.openai.com/introducing-openai/ |website=OpenAI Blog |date=December 12, 2015 |access-date=September 29, 2018 |archive-date=February 24, 2019 |archive-url=https://web.archive.org/web/20190224031626/https://blog.openai.com/introducing-openai/ |url-status=live }}{{cite news |date=December 12, 2015 |title=Tech giants pledge $1bn for 'altruistic AI' venture, OpenAI |work=BBC News |url=https://www.bbc.com/news/technology-35082344 |url-status=live |access-date=December 19, 2015 |archive-url=https://web.archive.org/web/20180314021831/http://www.bbc.com/news/technology-35082344 |archive-date=March 14, 2018}} OpenAI was initially run from Brockman's living room.{{Cite news |last=Seetharaman |first=Deepa |date=September 27, 2024 |title=Turning OpenAI Into a Real Business Is Tearing It Apart |url=https://www.wsj.com/tech/ai/open-ai-division-for-profit-da26c24b |work=The Wall Street Journal |url-access=subscription |access-date=September 28, 2024 |archive-url=https://archive.today/20240928001622/https://www.wsj.com/tech/ai/open-ai-division-for-profit-da26c24b |archive-date=September 28, 2024 |url-status=live}} It was later headquartered at the Pioneer Building in the Mission District, San Francisco.{{Cite news|url=https://gizmodo.com/elon-musks-neuralink-sought-to-open-an-animal-testing-f-1823167674|title=Elon Musk's Neuralink Sought to Open an Animal Testing Facility in San Francisco |last=Conger|first=Kate|work=Gizmodo|access-date=October 11, 2018|language=en-US|archive-date=September 24, 2018|archive-url=https://web.archive.org/web/20180924145028/https://gizmodo.com/elon-musks-neuralink-sought-to-open-an-animal-testing-f-1823167674|url-status=live}}{{Cite web|url=https://www.technologyreview.com/s/615181/ai-openai-moonshot-elon-musk-sam-altman-greg-brockman-messy-secretive-reality/|title=The messy, secretive reality behind OpenAI's bid to save the world |last=Hao |first=Karen |date=February 17, 2020|website=MIT Technology Review|language=en-US|access-date=March 9, 2020|archive-date=April 3, 2020 |archive-url=https://web.archive.org/web/20200403023123/https://www.technologyreview.com/s/615181/ai-openai-moonshot-elon-musk-sam-altman-greg-brockman-messy-secretive-reality/|url-status=live}}

According to OpenAI's charter, its founding mission is "to ensure that artificial general intelligence (AGI)—by which we mean highly autonomous systems that outperform humans at most economically valuable work—benefits all of humanity."

According to Wired, Brockman met with Yoshua Bengio, one of the "founding fathers" of deep learning, and drew up a list of the "best researchers in the field".{{cite magazine |author1=Cade Metz |date=April 27, 2016 |title=Inside OpenAI, Elon Musk's Wild Plan to Set Artificial Intelligence Free |url=https://www.wired.com/2016/04/openai-elon-musk-sam-altman-plan-to-set-artificial-intelligence-free/ |magazine=Wired |language=en-US |access-date=April 28, 2016 |archive-date=April 27, 2016 |archive-url=https://web.archive.org/web/20160427162700/http://www.wired.com/2016/04/openai-elon-musk-sam-altman-plan-to-set-artificial-intelligence-free/ |url-status=live }} Brockman was able to hire nine of them as the first employees in December 2015. In 2016, OpenAI paid corporate-level (rather than nonprofit-level) salaries, but did not pay AI researchers salaries comparable to those of Facebook or Google.

Microsoft's Peter Lee stated that the cost of a top AI researcher exceeds the cost of a top NFL quarterback prospect. OpenAI's potential and mission drew these researchers to the firm; a Google employee said he was willing to leave Google for OpenAI "partly because of the very strong group of people and, to a very large extent, because of its mission." Brockman stated that "the best thing that I could imagine doing was moving humanity closer to building real AI in a safe way." OpenAI co-founder Wojciech Zaremba stated that he turned down "borderline crazy" offers of two to three times his market value to join OpenAI instead.

In April 2016, OpenAI released a public beta of "OpenAI Gym", its platform for reinforcement learning research. Nvidia gifted its first DGX-1 supercomputer to OpenAI in August 2016 to help it train larger and more complex AI models with the capability of reducing processing time from six days to two hours.{{Cite news |last1=Carr |first1=Austin |last2=King |first2=Ian |date=June 15, 2023 |title=How Nvidia Became ChatGPT's Brain and Joined the $1 Trillion Club |url=https://www.bloomberg.com/news/features/2023-06-15/nvidia-s-ai-chips-power-chatgpt-and-multibillion-dollar-surge |publisher=Bloomberg News |url-access=subscription |archive-url=https://archive.today/20230618072913/https://www.bloomberg.com/news/features/2023-06-15/nvidia-s-ai-chips-power-chatgpt-and-multibillion-dollar-surge |archive-date=June 18, 2023 |url-status=live}}{{Cite news |last=Vanian |first=Jonathan |date=August 15, 2016 |title=Elon Musk's Artificial Intelligence Project Just Got a Free Supercomputer |url=https://fortune.com/2016/08/15/elon-musk-artificial-intelligence-openai-nvidia-supercomputer/ |work=Fortune |url-access=subscription |archive-url=https://archive.today/20230607233501/https://fortune.com/2016/08/15/elon-musk-artificial-intelligence-openai-nvidia-supercomputer/ |archive-date=June 7, 2023 |url-status=live}} In December 2016, OpenAI released "Universe", a software platform for measuring and training an AI's general intelligence across the world's supply of games, websites, and other applications.{{cite magazine|url=https://www.wired.com/2016/12/openais-universe-computers-learn-use-apps-like-humans/|title=Elon Musk's Lab Wants to Teach Computers to Use Apps Just Like Humans Do |last1=Metz |first1=Cade |magazine=WIRED |access-date=December 31, 2016|archive-date=January 13, 2019|archive-url=https://web.archive.org/web/20190113173007/https://www.wired.com/2016/12/openais-universe-computers-learn-use-apps-like-humans/|url-status=live}}{{cite news|url=https://techcrunch.com/2016/12/05/openais-universe-is-the-fun-parent-every-artificial-intelligence-deserves/|title=OpenAI's Universe is the fun parent every artificial intelligence deserves|last1=Mannes|first1=John|work=TechCrunch|access-date=December 31, 2016|archive-date=February 19, 2019|archive-url=https://web.archive.org/web/20190219080426/https://techcrunch.com/2016/12/05/openais-universe-is-the-fun-parent-every-artificial-intelligence-deserves/|url-status=live}}{{cite web|url=https://universe.openai.com/|title=OpenAI – Universe|language=en-us|access-date=December 31, 2016|archive-date=January 1, 2017|archive-url=https://web.archive.org/web/20170101001551/https://universe.openai.com/|url-status=dead}}{{cite web|url=https://www.theregister.co.uk/2016/12/05/openai_universe_reinforcement_learning/|title=Elon Musk-backed OpenAI reveals Universe – a universal training ground for computers|last1=Claburn|first1=Thomas|website=The Register|access-date=December 31, 2016|archive-date=January 1, 2017|archive-url=https://web.archive.org/web/20170101002022/http://www.theregister.co.uk/2016/12/05/openai_universe_reinforcement_learning/|url-status=live}}

In 2017, OpenAI spent $7.9 million, or a quarter of its functional expenses, on cloud computing alone.{{cite news |date=July 22, 2019 |title=Microsoft to invest $1 billion in OpenAI |language=en |work=Reuters |url=https://www.reuters.com/article/us-microsoft-openai/microsoft-to-invest-1-billion-in-openai-idUSKCN1UH1H9 |access-date=March 6, 2020 |archive-date=May 25, 2020 |archive-url=https://web.archive.org/web/20200525132055/https://www.reuters.com/article/us-microsoft-openai/microsoft-to-invest-1-billion-in-openai-idUSKCN1UH1H9 |url-status=live }} In comparison, DeepMind's total expenses in 2017 were $442 million. In the summer of 2018, simply training OpenAI's Dota 2 bots required renting 128,000 CPUs and 256 GPUs from Google for multiple weeks.

In 2018, Musk resigned from his Board of Directors seat, citing "a potential future conflict [of interest]" with his role as CEO of Tesla due to Tesla's AI development for self-driving cars.{{Cite web|url=https://www.theverge.com/2018/2/21/17036214/elon-musk-openai-ai-safety-leaves-board|title=Elon Musk leaves board of AI safety group to avoid conflict of interest with Tesla|first=James|last=Vincent|date=February 21, 2018|website=The Verge|access-date=February 22, 2018|archive-date=November 9, 2020|archive-url=https://web.archive.org/web/20201109015711/https://www.theverge.com/2018/2/21/17036214/elon-musk-openai-ai-safety-leaves-board|url-status=live}} Sam Altman claims that Musk believed that OpenAI had fallen behind other players like Google and Musk proposed instead to take over OpenAI himself, which the board rejected. Musk subsequently left OpenAI.

In February 2019, GPT-2 was announced, which gained attention for its ability to generate human-like text.{{cite web

|url = https://www.theguardian.com/technology/2019/feb/14/elon-musk-backed-ai-writes-convincing-news-fiction

|title = New AI fake text generator may be too dangerous to release, say creators

|last = Hern

|first = Alex

|work = The Guardian

|date = February 14, 2019

|access-date = December 19, 2020

|quote =

|archive-date = February 14, 2019

|archive-url = https://web.archive.org/web/20190214173112/https://www.theguardian.com/technology/2019/feb/14/elon-musk-backed-ai-writes-convincing-news-fiction

|url-status = live

}}

In 2019, OpenAI transitioned from non-profit to "capped" for-profit, with the profit being capped at 100 times any investment.{{Cite web |date=March 11, 2019 |title=OpenAI shifts from nonprofit to 'capped-profit' to attract capital |url=https://techcrunch.com/2019/03/11/openai-shifts-from-nonprofit-to-capped-profit-to-attract-capital/ |access-date=January 4, 2023 |language=en |archive-date=January 4, 2023 |archive-url=https://web.archive.org/web/20230104154138/https://techcrunch.com/2019/03/11/openai-shifts-from-nonprofit-to-capped-profit-to-attract-capital/ |url-status=live }} According to OpenAI, the capped-profit model allows OpenAI Global, LLC to legally attract investment from venture funds and, in addition, to grant employees stakes in the company.{{cite magazine |date=December 3, 2019 |title=To Compete With Google, OpenAI Seeks Investors–and Profits |url=https://www.wired.com/story/compete-google-openai-seeks-investorsand-profits/ |magazine=Wired |language=en |access-date=March 6, 2020 |archive-date=March 14, 2020 |archive-url=https://web.archive.org/web/20200314180028/https://www.wired.com/story/compete-google-openai-seeks-investorsand-profits/ |url-status=live }} Many top researchers work for Google Brain, DeepMind, or Facebook, which offer stock options that a nonprofit would be unable to.{{cite news |last1=Kahn |first1=Jeremy |date=March 11, 2019 |title=AI Research Group Co-Founded by Elon Musk Starts For-Profit Arm |work=Bloomberg News |url=https://www.bloomberg.com/news/articles/2019-03-11/ai-research-group-co-founded-by-musk-starts-for-profit-arm |access-date=March 6, 2020 |archive-date=December 7, 2019 |archive-url=https://web.archive.org/web/20191207080100/https://www.bloomberg.com/news/articles/2019-03-11/ai-research-group-co-founded-by-musk-starts-for-profit-arm |url-status=live }} Before the transition, public disclosure of the compensation of top employees at OpenAI was legally required.{{Cite news |last=Metz |first=Cade |date=April 19, 2018 |title=A.I. Researchers Are Making More Than $1 Million, Even at a Nonprofit |language=en-US |work=The New York Times |url=https://www.nytimes.com/2018/04/19/technology/artificial-intelligence-salaries-openai.html |access-date=January 28, 2023 |issn=0362-4331 |archive-date=August 8, 2018 |archive-url=https://web.archive.org/web/20180808200610/https://www.nytimes.com/2018/04/19/technology/artificial-intelligence-salaries-openai.html |url-status=live }}

The company then distributed equity to its employees and partnered with Microsoft,{{Cite web |date=July 22, 2019 |title=Microsoft invests in and partners with OpenAI |url=https://openai.com/blog/microsoft-invests-in-and-partners-with-openai |language=en |access-date=March 17, 2023 |archive-date=February 28, 2023 |archive-url=https://web.archive.org/web/20230228125554/https://openai.com/blog/microsoft-invests-in-and-partners-with-openai |url-status=live }} announcing an investment package of $1 billion into the company. Since then, OpenAI systems have run on an Azure-based supercomputing platform from Microsoft.{{cite web | last=Langston | first=Jennifer | title=Microsoft announces new supercomputer, lays out vision for future AI work | website=Source | date=January 11, 2023 | quote=Built in collaboration with and exclusively for OpenAI | url=https://news.microsoft.com/source/features/ai/openai-azure-supercomputer/ | archive-url=https://web.archive.org/web/20230210180449/https://news.microsoft.com/source/features/ai/openai-azure-supercomputer/ | archive-date=February 10, 2023 | url-status=live | access-date=February 10, 2023 }}{{cite web | last=Foley | first=Mary Jo | title=Microsoft builds a supercomputer for OpenAI for training massive AI models | website=ZDNET | date=May 19, 2020 | quote= | url=https://www.zdnet.com/article/microsoft-builds-a-supercomputer-for-openai-for-training-massive-ai-models/ | archive-url=https://web.archive.org/web/20230210180846/https://www.zdnet.com/article/microsoft-builds-a-supercomputer-for-openai-for-training-massive-ai-models/ | archive-date=February 10, 2023 | url-status=live | access-date=February 10, 2023 }}{{cite web | title=Microsoft's OpenAI supercomputer has 285,000 CPU cores, 10,000 GPUs | quote=Microsoft's OpenAI supercomputer has 285,000 CPU cores, 10,000 GPUs. It's one of the five fastest systems in the world. | website=Engadget | date=May 19, 2020 | url=https://www.engadget.com/microsoft-openai-supercomputer-azure-150001119.html | archive-url=https://web.archive.org/web/20230210180859/https://www.engadget.com/microsoft-openai-supercomputer-azure-150001119.html | archive-date=February 10, 2023 | url-status=live | access-date=February 10, 2023 }}

OpenAI Global, LLC then announced its intention to commercially license its technologies.{{Cite web |date=July 22, 2019 |title=Microsoft Invests in and Partners with OpenAI to Support Us Building Beneficial AGI |url=https://openai.com/blog/microsoft/ |access-date=February 21, 2020 |website=OpenAI |language=en |archive-date=November 7, 2020 |archive-url=https://web.archive.org/web/20201107230518/https://openai.com/blog/microsoft/ |url-status=live }} It planned to spend the $1 billion "within five years, and possibly much faster".{{cite news |last1=Murgia |first1=Madhumita |date=August 7, 2019 |title=DeepMind runs up higher losses and debts in race for AI |work=Financial Times |url=https://www.ft.com/content/d4280856-b92d-11e9-8a88-aa6628ac896c |access-date=March 6, 2020 |archive-date=December 26, 2019 |archive-url=https://web.archive.org/web/20191226185727/https://www.ft.com/content/d4280856-b92d-11e9-8a88-aa6628ac896c |url-status=live }} Altman has stated that even a billion dollars may turn out to be insufficient, and that the lab may ultimately need "more capital than any non-profit has ever raised" to achieve artificial general intelligence.{{cite news |title=OpenAI Will Need More Capital Than Any Non-Profit Has Ever Raised |language=en |work=Fortune |url=https://fortune.com/2019/10/03/openai-will-need-more-capital-than-any-non-profit-has-ever-raised/ |access-date=March 6, 2020 |archive-date=December 8, 2019 |archive-url=https://web.archive.org/web/20191208040513/https://fortune.com/2019/10/03/openai-will-need-more-capital-than-any-non-profit-has-ever-raised/ |url-status=live }}

The transition from a nonprofit to a capped-profit company was viewed with skepticism by Oren Etzioni of the nonprofit Allen Institute for AI, who agreed that wooing top researchers to a nonprofit is difficult, but stated "I disagree with the notion that a nonprofit can't compete" and pointed to successful low-budget projects by OpenAI and others. "If bigger and better funded was always better, then IBM would still be number one."

The nonprofit, OpenAI, Inc., is the sole controlling shareholder of OpenAI Global, LLC, which, despite being a for-profit company, retains a formal fiduciary responsibility to OpenAI, Inc.'s nonprofit charter. A majority of OpenAI, Inc.'s board is barred from having financial stakes in OpenAI Global, LLC. In addition, minority members with a stake in OpenAI Global, LLC are barred from certain votes due to conflict of interest. Some researchers have argued that OpenAI Global, LLC's switch to for-profit status is inconsistent with OpenAI's claims to be "democratizing" AI.{{cite news |last1=Vincent |first1=James |date=July 22, 2019 |title=Microsoft invests $1 billion in OpenAI to pursue holy grail of artificial intelligence |language=en |work=The Verge |url=https://www.theverge.com/2019/7/22/20703578/microsoft-openai-investment-partnership-1-billion-azure-artificial-general-intelligence-agi |access-date=March 6, 2020 |archive-date=July 23, 2019 |archive-url=https://web.archive.org/web/20190723011910/https://www.theverge.com/2019/7/22/20703578/microsoft-openai-investment-partnership-1-billion-azure-artificial-general-intelligence-agi |url-status=live }}

In 2020, OpenAI announced GPT-3, a language model trained on large internet datasets. GPT-3 is aimed at natural language answering questions, but it can also translate between languages and coherently generate improvised text. It also announced that an associated API, named simply "the API", would form the heart of its first commercial product.{{cite news |url=https://www.bloomberg.com/news/articles/2020-06-11/trillions-of-words-analyzed-openai-sets-loose-ai-language-colossus |title=Trillions of Words Analyzed, OpenAI Sets Loose AI Language Colossus |last=Vance |first=Ashlee |author-link=Ashlee Vance |date=June 11, 2020 |work=Bloomberg News |access-date=June 12, 2020 |archive-date=October 13, 2020 |archive-url=https://web.archive.org/web/20201013163856/https://www.bloomberg.com/news/articles/2020-06-11/trillions-of-words-analyzed-openai-sets-loose-ai-language-colossus |url-status=live }}

Eleven employees left OpenAI, mostly between December 2020 and January 2021, in order to establish Anthropic.{{Cite news |last=Moss |first=Sebastian |date=June 2, 2021 |title=Eleven OpenAI Employees Break Off to Establish Anthropic, Raise $124 Million |url=https://aibusiness.com/verticals/eleven-openai-employees-break-off-to-establish-anthropic-raise-124m |work=AI Business |access-date=June 24, 2024 |archive-date=June 6, 2024 |archive-url=https://web.archive.org/web/20240606002642/https://aibusiness.com/verticals/eleven-openai-employees-break-off-to-establish-anthropic-raise-124m |url-status=live }}

In 2021, OpenAI introduced DALL-E, a specialized deep learning model adept at generating complex digital images from textual descriptions, utilizing a variant of the GPT-3 architecture.{{cite web |url=https://venturebeat.com/2021/01/05/openai-debuts-dall-e-for-generating-images-from-text/ |title=OpenAI debuts DALL-E for generating images from text |publisher=VentureBeat |date=January 5, 2021 |access-date=January 5, 2021 |archive-date=January 5, 2021 |archive-url=https://web.archive.org/web/20210105221534/https://venturebeat.com/2021/01/05/openai-debuts-dall-e-for-generating-images-from-text/ }}

File:UK national football team considering compete in UEFA Euro and FIFA World Cup – ChatGPT.jpg was a major event in the AI boom. By January 2023, ChatGPT had become what was then the fastest-growing consumer software application in history, gaining over 100 million users in two months.{{cite web |last1=Porter |first1=Jon |title=ChatGPT continues to be one of the fastest-growing services ever |url=https://www.theverge.com/2023/11/6/23948386/chatgpt-active-user-count-openai-developer-conference |website=The Verge |access-date=26 November 2024 |language=en |date=6 November 2023 |archive-date=November 27, 2024 |archive-url=https://web.archive.org/web/20241127073502/https://www.theverge.com/2023/11/6/23948386/chatgpt-active-user-count-openai-developer-conference |url-status=live }}]]

In December 2022, OpenAI received widespread media coverage after launching a free preview of ChatGPT, its new AI chatbot based on GPT-3.5. According to OpenAI, the preview received over a million signups within the first five days.{{cite news |last1=Roose |first1=Kevin |title=The Brilliance and Weirdness of ChatGPT |url=https://www.nytimes.com/2022/12/05/technology/chatgpt-ai-twitter.html |access-date=January 5, 2023 |work=The New York Times |date=December 5, 2022 |archive-date=January 18, 2023 |archive-url=https://web.archive.org/web/20230118134332/https://www.nytimes.com/2022/12/05/technology/chatgpt-ai-twitter.html |url-status=live }} According to anonymous sources cited by Reuters in December 2022, OpenAI Global, LLC was projecting $200 million of revenue in 2023 and $1 billion in revenue in 2024.{{cite news |last1=Dastin |first1=Jeffrey |last2=Hu |first2=Krystal |last3=Dave |first3=Paresh |last4=Dave |first4=Paresh |title=Exclusive: ChatGPT owner OpenAI projects $1 billion in revenue by 2024 |url=https://www.reuters.com/business/chatgpt-owner-openai-projects-1-billion-revenue-by-2024-sources-2022-12-15/ |access-date=January 5, 2023 |work=Reuters |date=December 15, 2022 |language=en |archive-date=February 3, 2023 |archive-url=https://web.archive.org/web/20230203201121/https://www.reuters.com/business/chatgpt-owner-openai-projects-1-billion-revenue-by-2024-sources-2022-12-15/ |url-status=live }}

In January 2023, OpenAI Global, LLC was in talks for funding that would value the company at $29 billion, double its 2021 value.{{Cite news |last=Kruppa |first=Berber Jin and Miles |title=WSJ News Exclusive {{!}} ChatGPT Creator in Investor Talks at $29 Billion Valuation |url=https://www.wsj.com/articles/chatgpt-creator-openai-is-in-talks-for-tender-offer-that-would-value-it-at-29-billion-11672949279 |access-date=January 6, 2023 |newspaper=Wall Street Journal |date=January 5, 2023 |language=en-US |archive-date=February 3, 2023 |archive-url=https://web.archive.org/web/20230203201104/https://www.wsj.com/articles/chatgpt-creator-openai-is-in-talks-for-tender-offer-that-would-value-it-at-29-billion-11672949279 |url-status=live }} On January 23, 2023, Microsoft announced a new US$10 billion investment in OpenAI Global, LLC over multiple years, partially needed to use Microsoft's cloud-computing service Azure.{{Cite news |date=January 23, 2023 |title=Microsoft Adds $10 Billion to Investment in ChatGPT Maker OpenAI |language=en |work=Bloomberg.com |url=https://www.bloomberg.com/news/articles/2023-01-23/microsoft-makes-multibillion-dollar-investment-in-openai |access-date=January 23, 2023 |archive-date=January 23, 2023 |archive-url=https://web.archive.org/web/20230123180447/https://www.bloomberg.com/news/articles/2023-01-23/microsoft-makes-multibillion-dollar-investment-in-openai |url-status=live }}{{Cite web |last=Capoot |first=Ashley |title=Microsoft announces multibillion-dollar investment in ChatGPT-maker OpenAI |url=https://www.cnbc.com/2023/01/23/microsoft-announces-multibillion-dollar-investment-in-chatgpt-maker-openai.html |access-date=January 23, 2023 |website=CNBC |date=January 23, 2023 |language=en |archive-date=January 23, 2023 |archive-url=https://web.archive.org/web/20230123154020/https://www.cnbc.com/2023/01/23/microsoft-announces-multibillion-dollar-investment-in-chatgpt-maker-openai.html |url-status=live }} Rumors of this deal suggested that Microsoft may receive 75% of OpenAI's profits until it secures its investment return and a 49% stake in the company.{{Cite web |last=Warren |first=Tom |date=January 23, 2023 |title=Microsoft extends OpenAI partnership in a "multibillion dollar investment" |url=https://www.theverge.com/2023/1/23/23567448/microsoft-openai-partnership-extension-ai |access-date=April 29, 2023 |website=The Verge |language=en-US |archive-date=April 29, 2023 |archive-url=https://web.archive.org/web/20230429020214/https://www.theverge.com/2023/1/23/23567448/microsoft-openai-partnership-extension-ai |url-status=live }} The investment is believed to be a part of Microsoft's efforts to integrate OpenAI's ChatGPT into the Bing search engine. Google announced a similar AI application (Bard), after ChatGPT was launched, fearing that ChatGPT could threaten Google's place as a go-to source for information.{{Cite news |date=February 6, 2023 |title=Bard: Google launches ChatGPT rival |language=en-GB |work=BBC News |url=https://www.bbc.com/news/technology-64546299 |access-date=February 7, 2023 |archive-date=February 7, 2023 |archive-url=https://web.archive.org/web/20230207111531/https://www.bbc.com/news/technology-64546299 |url-status=live }}{{Cite web |last=Vincent |first=James |date=February 8, 2023 |title=Google's AI chatbot Bard makes factual error in first demo |url=https://www.theverge.com/2023/2/8/23590864/google-ai-chatbot-bard-mistake-error-exoplanet-demo |access-date=February 12, 2023 |website=The Verge |language=en-US |archive-date=February 12, 2023 |archive-url=https://web.archive.org/web/20230212094317/https://www.theverge.com/2023/2/8/23590864/google-ai-chatbot-bard-mistake-error-exoplanet-demo |url-status=live }}

On February 7, 2023, Microsoft announced that it was building AI technology based on the same foundation as ChatGPT into Microsoft Bing, Edge, Microsoft 365 and other products.{{Cite news |last=Dotan |first=Tom |title=Microsoft Adds ChatGPT AI Technology to Bing Search Engine |url=https://www.wsj.com/articles/microsoft-adds-chatgpt-ai-technology-to-bing-search-engine-11675793525 |access-date=February 7, 2023 |newspaper=Wall Street Journal |date=February 7, 2023 |language=en-US |archive-date=February 7, 2023 |archive-url=https://web.archive.org/web/20230207185720/https://www.wsj.com/articles/microsoft-adds-chatgpt-ai-technology-to-bing-search-engine-11675793525 |url-status=live }}

On March 3, 2023, Reid Hoffman resigned from his board seat, citing a desire to avoid conflicts of interest with his investments in AI companies via Greylock Partners, and his co-founding of the AI startup Inflection AI. Hoffman remained on the board of Microsoft, a major investor in OpenAI.{{Cite news |last=Dastin |first=Jeffrey |date=March 3, 2023 |title=OpenAI's long-time backer Reid Hoffman leaves board |language=en |work=Reuters |url=https://www.reuters.com/technology/openais-long-time-backer-reid-hoffman-leaves-board-2023-03-03/ |access-date=March 17, 2023 |archive-date=May 13, 2023 |archive-url=https://web.archive.org/web/20230513191651/https://www.reuters.com/technology/openais-long-time-backer-reid-hoffman-leaves-board-2023-03-03/ |url-status=live }}

On March 14, 2023, OpenAI released GPT-4, both as an API (with a waitlist) and as a feature of ChatGPT Plus.{{Cite web |title=GPT-4 |url=https://openai.com/product/gpt-4 |access-date=March 16, 2023 |website=openai.com |language=en-US |archive-date=March 14, 2023 |archive-url=https://web.archive.org/web/20230314165712/https://openai.com/product/gpt-4 |url-status=live }}

On May 22, 2023, Sam Altman, Greg Brockman and Ilya Sutskever posted recommendations for the governance of superintelligence.{{Cite web |title=Governance of superintelligence |url=https://openai.com/blog/governance-of-superintelligence |access-date=May 30, 2023 |website=openai.com |language=en-US |archive-date=May 27, 2023 |archive-url=https://web.archive.org/web/20230527061619/https://openai.com/blog/governance-of-superintelligence |url-status=live }} They consider that superintelligence could happen within the next 10 years, allowing a "dramatically more prosperous future" and that "given the possibility of existential risk, we can't just be reactive". They propose creating an international watchdog organization similar to IAEA to oversee AI systems above a certain capability threshold, suggesting that relatively weak AI systems on the other side should not be overly regulated. They also call for more technical safety research for superintelligences, and ask for more coordination, for example through governments launching a joint project which "many current efforts become part of".{{Cite web |last1=Wodecki |first1=Ben |last2=Yao |first2=Deborah |date=May 23, 2023 |title=OpenAI Founders Warn AI 'Superintelligence' is Like Nuclear Power |url=https://aibusiness.com/responsible-ai/openai-leaders-want-the-public-to-decide-ai-rules |access-date=May 30, 2023 |archive-date=May 30, 2023 |archive-url=https://web.archive.org/web/20230530042308/https://aibusiness.com/responsible-ai/openai-leaders-want-the-public-to-decide-ai-rules |url-status=live }}

In July 2023, OpenAI launched the superalignment project, aiming to find within 4 years how to align future superintelligences by automating alignment research using AI.{{Cite web |date=July 5, 2023 |title=Introducing Superalignment |url=https://openai.com/index/introducing-superalignment/ |website=OpenAI |access-date=May 25, 2024 |archive-date=May 25, 2024 |archive-url=https://web.archive.org/web/20240525041645/https://openai.com/index/introducing-superalignment/ |url-status=live }}

In August 2023, it was announced that OpenAI had acquired the New York-based start-up Global Illumination, a company that deploys AI to develop digital infrastructure and creative tools.{{Cite news |date=August 16, 2023 |title=OpenAI acquires start-up Global Illumination to work on core products, ChatGPT |language=en |work=Reuters |url=https://www.reuters.com/markets/deals/openai-acquires-start-up-global-illumination-work-core-products-chatgpt-2023-08-16/ |access-date=August 17, 2023 |archive-date=August 17, 2023 |archive-url=https://web.archive.org/web/20230817111738/https://www.reuters.com/markets/deals/openai-acquires-start-up-global-illumination-work-core-products-chatgpt-2023-08-16/ |url-status=live }}

On September 21, 2023, Microsoft had begun rebranding all variants of its Copilot to Microsoft Copilot, including the former Bing Chat and the Microsoft 365 Copilot.{{Cite web |last=Edwards |first=Nathan |date=September 21, 2023 |title=Microsoft's unified Copilot is coming to Windows, Edge, and everywhere else |url=https://www.theverge.com/2023/9/21/23883798/microsoft-copilot-unified-windows-11-apps-launch-date |url-status=live |archive-url=https://web.archive.org/web/20231207200142/https://www.theverge.com/2023/9/21/23883798/microsoft-copilot-unified-windows-11-apps-launch-date |archive-date=December 7, 2023 |access-date=February 2, 2024 |website=The Verge |language=en}} This strategy was followed in December 2023 by adding the MS-Copilot to many installations of Windows 11 and Windows 10 as well as a standalone Microsoft Copilot app released for Android{{Cite web |last=Warren |first=Tom |date=December 26, 2023 |title=Microsoft Copilot is now available as a ChatGPT-like app on Android |url=https://www.theverge.com/2023/12/26/24015198/microsoft-copilot-mobile-app-android-launch |url-status=live |archive-url=https://web.archive.org/web/20240131115238/https://www.theverge.com/2023/12/26/24015198/microsoft-copilot-mobile-app-android-launch |archive-date=January 31, 2024 |access-date=February 2, 2024 |website=The Verge |language=en}} and one released for iOS thereafter.{{Cite web |date=December 29, 2023 |title=Microsoft's Copilot app is now available on iOS |url=https://www.theverge.com/2023/12/29/24019288/microsoft-copilot-app-available-iphone-ipad-ai |url-status=live |archive-url=https://web.archive.org/web/20240130064608/https://www.theverge.com/2023/12/29/24019288/microsoft-copilot-app-available-iphone-ipad-ai |archive-date=January 30, 2024 |access-date=February 2, 2024 |website=The Verge |publisher=Vox Media}}

In October 2023, Sam Altman and Peng Xiao, CEO of the Emirati AI firm G42, announced Open AI would let G42 deploy Open AI technology.{{Cite web |last=Northrop |first=Katrina |date=December 4, 2023 |title=G42's Ties To China Run Deep |url=https://www.thewirechina.com/2023/12/03/g42s-ties-to-china-run-deep-g42-peng-xiao/ |url-status=live |archive-url=https://web.archive.org/web/20240125203013/https://www.thewirechina.com/2023/12/03/g42s-ties-to-china-run-deep-g42-peng-xiao/ |archive-date=January 25, 2024 |access-date=February 17, 2024 |website=The Wire China |language=en-US}}

On November 6, 2023, OpenAI launched GPTs, allowing individuals to create customized versions of ChatGPT for specific purposes, further expanding the possibilities of AI applications across various industries.{{Cite news |date=November 6, 2023 |title=OpenAI Launches Custom ChatGPT Versions |language=en |work=The New York Times |url=https://www.nytimes.com/2023/11/06/technology/openai-custom-chatgpt.html |access-date=November 12, 2023 |archive-date=November 12, 2023 |archive-url=https://web.archive.org/web/20231112122229/https://www.nytimes.com/2023/11/06/technology/openai-custom-chatgpt.html |url-status=live }} On November 14, 2023, OpenAI announced they temporarily suspended new sign-ups for ChatGPT Plus due to high demand.{{Cite news |last=Elstrom |first=Peter |date=November 15, 2023 |title=OpenAI Pauses New Signups to Manage Overwhelming Demand |language=en |work=Bloomberg |url=https://www.bloomberg.com/news/articles/2023-11-15/openai-pauses-new-signups-to-manage-overwhelming-demand |url-status=live |access-date=November 15, 2023 |archive-url=https://web.archive.org/web/20231115130109/https://www.bloomberg.com/news/articles/2023-11-15/openai-pauses-new-signups-to-manage-overwhelming-demand |archive-date=November 15, 2023}} Access for newer subscribers re-opened a month later on December 13.{{cite web |last1=Idris |first1=Abubakar |title=OpenAI Reopens ChatGPT Plus Subscriptions |url=https://themessenger.com/tech/openai-re-opens-chatgpt-plus-subscriptions |website=The Messenger |language=en |date=December 13, 2023 |access-date=December 14, 2023 |archive-date=December 14, 2023 |archive-url=https://web.archive.org/web/20231214020417/https://themessenger.com/tech/openai-re-opens-chatgpt-plus-subscriptions |url-status=dead }}

In January 2024, OpenAI partnered with Arizona State University to provide complete access to ChatGPT Enterprise in its first educational collaboration.{{Cite web |last=Field |first=Hayden |date=January 18, 2024 |title=OpenAI announces first partnership with a university |url=https://www.cnbc.com/2024/01/18/openai-announces-first-partnership-with-a-university.html |access-date=January 18, 2024 |website=CNBC |language=en |archive-date=January 24, 2024 |archive-url=https://web.archive.org/web/20240124135634/https://www.cnbc.com/2024/01/18/openai-announces-first-partnership-with-a-university.html |url-status=live }}

In February, amidst SEC probes and investigations into CEO Altman's communications{{cite news |last=Seetharaman |first=Deepa |date=February 28, 2024 |title=SEC Investigating Whether OpenAI Investors Were Misled |url=https://www.wsj.com/tech/sec-investigating-whether-openai-investors-were-misled-9d90b411 |url-status=live |archive-url=https://web.archive.org/web/20240229044238/https://www.wsj.com/tech/sec-investigating-whether-openai-investors-were-misled-9d90b411 |archive-date=February 29, 2024 |access-date=February 29, 2024 |work=The Wall Street Journal |publisher=News Corp}} OpenAI unveiled its text-to-video model Sora (text-to-video model), currently available to red teams for managing risks{{cite web |last1=Metz |first1=Cade |title=OpenAI Unveils A.I. That Instantly Generates Eye-Popping Videos |url=https://www.nytimes.com/2024/02/15/technology/openai-sora-videos.html |website=The New York Times |date=February 15, 2024 |access-date=February 16, 2024 |archive-date=February 15, 2024 |archive-url=https://web.archive.org/web/20240215220626/https://www.nytimes.com/2024/02/15/technology/openai-sora-videos.html |url-status=live }}{{Cite web |title=Sora: Creating video from text |url=https://openai.com/sora |access-date=February 17, 2024 |website=openai.com |language=en-US |archive-date=February 17, 2024 |archive-url=https://web.archive.org/web/20240217015001/https://openai.com/sora |url-status=live }}

On February 29, 2024, Elon Musk filed a lawsuit against OpenAI and CEO Sam Altman, accusing them of shifting focus from public benefit to profit maximization—a case OpenAI dismissed as “incoherent” and “frivolous,” though Musk later revived legal action against Altman and others in August 2024.{{Cite news |last1=Satariano |first1=Adam |last2=Metz |first2=Cade |last3=Mickle |first3=Tripp |date=March 1, 2024 |title=Elon Musk Sues OpenAI and Sam Altman for Violating the Company's Principles |url=https://www.nytimes.com/2024/03/01/technology/elon-musk-openai-sam-altman-lawsuit.html |access-date=March 2, 2024 |work=The New York Times |language=en-US |issn=0362-4331 |archive-date=March 6, 2024 |archive-url=https://web.archive.org/web/20240306201120/https://www.nytimes.com/2024/03/01/technology/elon-musk-openai-sam-altman-lawsuit.html |url-status=live }}{{Cite web |last=Lopatto |first=Elizabeth |date=March 6, 2024 |title=OpenAI says Elon Musk wanted 'absolute control' of the company |url=https://www.theverge.com/2024/3/5/24091773/openai-response-elon-musk-breach-of-contract-lawsuit |access-date=March 6, 2024 |website=The Verge |language=en |archive-date=March 10, 2024 |archive-url=https://web.archive.org/web/20240310050448/https://www.theverge.com/2024/3/5/24091773/openai-response-elon-musk-breach-of-contract-lawsuit |url-status=live }}{{cite news|last=Kharpal|first=Arjun|date=August 5, 2024|title=Elon Musk revives lawsuit against OpenAI, Sam Altman in federal court|publisher=CNBC|url=https://www.cnbc.com/2024/08/05/elon-musk-revives-lawsuit-against-openai-sam-altman-in-federal-court.html|access-date=August 5, 2024|archive-date=August 5, 2024|archive-url=https://web.archive.org/web/20240805125804/https://www.cnbc.com/2024/08/05/elon-musk-revives-lawsuit-against-openai-sam-altman-in-federal-court.html|url-status=live}}{{cite news|last=De Avila|first=Joseph|date=August 5, 2024|title=Elon Musk Revives Lawsuit Against OpenAI and Sam Altman|work=The Wall Street Journal|publisher=News Corp|url=https://www.wsj.com/tech/ai/elon-musk-revives-lawsuit-against-openai-and-sam-altman-d7e5a87c|access-date=August 5, 2024|archive-date=August 5, 2024|archive-url=https://web.archive.org/web/20240805154833/https://www.wsj.com/tech/ai/elon-musk-revives-lawsuit-against-openai-and-sam-altman-d7e5a87c|url-status=live}}

In May 2024, significant leadership changes occurred as Chief Scientist Ilya Sutskever resigned—being succeeded by Jakub Pachocki—and co-leader Jan Leike departed amid concerns over safety and trust.{{cite web |last1=Hollister |first1=Sean |title=OpenAI chief scientist Ilya Sutskever is officially leaving |url=https://www.theverge.com/2024/5/14/24156920/openai-chief-scientist-ilya-sutskever-leaves |website=The Verge |access-date=May 14, 2024 |language=en |date=May 14, 2024 |archive-date=May 14, 2024 |archive-url=https://web.archive.org/web/20240514234107/https://www.theverge.com/2024/5/14/24156920/openai-chief-scientist-ilya-sutskever-leaves |url-status=live }}{{Cite web |last=Samuel |first=Sigal |date=May 17, 2024 |title="I lost trust": Why the OpenAI team in charge of safeguarding humanity imploded |url=https://www.vox.com/future-perfect/2024/5/17/24158403/openai-resignations-ai-safety-ilya-sutskever-jan-leike-artificial-intelligence |access-date=May 18, 2024 |website=Vox |language=en |archive-date=May 18, 2024 |archive-url=https://web.archive.org/web/20240518205458/https://www.vox.com/future-perfect/2024/5/17/24158403/openai-resignations-ai-safety-ilya-sutskever-jan-leike-artificial-intelligence |url-status=live }} That same month, OpenAI formed a partnership with Reddit to integrate its content into OpenAI products{{Cite web |date=May 16, 2024 |title=Reddit and OpenAI Build Partnership - Upvoted |url=https://www.redditinc.com/blog/reddit-and-oai-partner |access-date=May 19, 2024 |website=www.redditinc.com |language=en-US |archive-date=May 19, 2024 |archive-url=https://web.archive.org/web/20240519012615/https://www.redditinc.com/blog/reddit-and-oai-partner |url-status=live }} and inked content deals with News Corp, along with licensing arrangements involving publishers such as Axios and Vox Media.{{Cite web |date=May 29, 2024 |title=Exclusive: The Atlantic, Vox Media ink licensing, product deals with OpenAI |url=https://www.axios.com/2024/05/29/atlantic-vox-media-openai-licensing-deal |access-date=June 3, 2024 |website=Axios |last=Fischer |first=Sara |archive-date=May 31, 2024 |archive-url=https://web.archive.org/web/20240531134558/https://www.axios.com/2024/05/29/atlantic-vox-media-openai-licensing-deal|url-status=live }}

In June 2024, OpenAI joined forces with Apple Inc. to integrate ChatGPT features into Apple Intelligence and iPhone{{Cite web |last=Wiggers |first=Kyle |date=June 10, 2024 |title=Apple brings ChatGPT to its apps, including Siri |url=https://techcrunch.com/2024/06/10/apple-brings-chatgpt-to-its-apps-including-siri/ |access-date=June 10, 2024 |website=TechCrunch |language=en-US |archive-date=June 10, 2024 |archive-url=https://web.archive.org/web/20240610190918/https://techcrunch.com/2024/06/10/apple-brings-chatgpt-to-its-apps-including-siri/ |url-status=live }} and added former NSA head Paul Nakasone to its board,{{cite web |last1=Coldewey |first1=Devin |title=Former NSA head joins OpenAI board and safety committee |url=https://techcrunch.com/2024/06/13/former-nsa-head-joins-openai-board-and-safety-committee/ |website=TechCrunch |access-date=June 15, 2024 |date=June 13, 2024 |archive-date=June 14, 2024 |archive-url=https://web.archive.org/web/20240614000140/https://techcrunch.com/2024/06/13/former-nsa-head-joins-openai-board-and-safety-committee/ |url-status=live }} while acquiring Multi, a startup focused on remote collaboration.{{cite web |last1=Wiggers |first1=Kyle |title=OpenAI buys a remote collaboration platform |url=https://techcrunch.com/2024/06/24/openai-buys-a-remote-collaboration-platform/ |website=TechCrunch |access-date=July 5, 2024 |date=June 24, 2024 |archive-date=June 29, 2024 |archive-url=https://web.archive.org/web/20240629064556/https://techcrunch.com/2024/06/24/openai-buys-a-remote-collaboration-platform/ |url-status=live }}

In July 2024, Reuters reported that OpenAI was developing a project, codenamed ‘Strawberry’, to enhance AI reasoning—a project later released in September as the o1 model.{{Cite news |last1=Tong |first1=Anna |last2=Paul |first2=Katie |date=July 15, 2024 |title=Exclusive: OpenAI working on new reasoning technology under code name 'Strawberry' |url=https://www.reuters.com/technology/artificial-intelligence/openai-working-new-reasoning-technology-under-code-name-strawberry-2024-07-12/ |access-date=July 22, 2024 |work=Reuters |archive-date=July 13, 2024 |archive-url=https://web.archive.org/web/20240713080951/https://www.reuters.com/technology/artificial-intelligence/openai-working-new-reasoning-technology-under-code-name-strawberry-2024-07-12/ |url-status=live }}{{Cite web |last=Sager |first=Monica |date=July 16, 2024 |title=What we know about OpenAI's secretive 'Project Strawberry' |url=https://www.newsweek.com/openai-strawberry-chat-gpt-ai-sam-altman-1925960 |access-date=July 22, 2024 |website=Newsweek |language=en |archive-date=July 22, 2024 |archive-url=https://web.archive.org/web/20240722214700/https://www.newsweek.com/openai-strawberry-chat-gpt-ai-sam-altman-1925960 |url-status=live }}

In August 2024, cofounder John Schulman left to join rival startup Anthropic, and OpenAI’s president Greg Brockman took extended leave until November.{{cite web |last=Wiggers |first=Kyle |date=August 6, 2024 |title=OpenAI co-founder Schulman leaves for Anthropic, Brockman takes extended leave |url=https://techcrunch.com/2024/08/05/openai-co-founder-leaves-for-anthropic/ |access-date=August 7, 2024 |website=TechCrunch |language=en-US |archive-date=August 7, 2024 |archive-url=https://web.archive.org/web/20240807070712/https://techcrunch.com/2024/08/05/openai-co-founder-leaves-for-anthropic/ |url-status=live }}

In September 2024, OpenAI’s global affairs chief endorsed the UK's “smart” AI regulation during testimony to a House of Lords committee,{{Cite web |date=2024-09-23 |title=OpenAI 'in favour' of UK AI legislation, policy chief says |url=https://www.uktech.news/ai/openai-in-favour-of-uk-ai-legislation-policy-chief-says-20240923 |access-date=2024-09-24 |website=UKTN |language=en-GB |archive-date=September 27, 2024 |archive-url=https://web.archive.org/web/20240927171401/https://www.uktech.news/ai/openai-in-favour-of-uk-ai-legislation-policy-chief-says-20240923 |url-status=live }} Meanwhile, CTO Mira Murati announced her departure amid internal concerns.{{Cite web |last=Field |first=Hayden |date=2024-09-25 |title=OpenAI CTO Mira Murati announces she's leaving the company |url=https://www.cnbc.com/2024/09/25/openai-cto-mira-murati-announces-shes-leaving-the-company.html |access-date=2024-09-25 |website=CNBC |language=en |archive-date=September 25, 2024 |archive-url=https://archive.today/20240925211351/https://www.cnbc.com/2024/09/25/openai-cto-mira-murati-announces-shes-leaving-the-company.html |url-status=live }}{{Cite web |author1=Maxwell Zeff |author2=Kyle Wiggers |date=2024-09-25 |title=OpenAI CTO Mira Murati says she's leaving the company |url=https://techcrunch.com/2024/09/25/openai-cto-mira-murati-says-shes-leaving-the-company/ |access-date=2024-09-27 |website=TechCrunch |language=en-US |archive-date=October 4, 2024 |archive-url=https://web.archive.org/web/20241004200114/https://techcrunch.com/2024/09/25/openai-cto-mira-murati-says-shes-leaving-the-company/ |url-status=live }}

In October 2024, OpenAI secured $6.6 billion in funding—valuing it at $157 billion—with major investors including Microsoft, Nvidia, and SoftBank,{{cite news |last=Hu |first=Krystal |title=OpenAI closes $6.6 billion funding haul with investment from Microsoft and Nvidia |date=2 October 2024 |url=https://www.reuters.com/technology/artificial-intelligence/openai-closes-66-billion-funding-haul-valuation-157-billion-with-investment-2024-10-02 |website=Reuters |access-date=3 October 2024 |archive-date=October 6, 2024 |archive-url=https://web.archive.org/web/20241006092108/https://www.reuters.com/technology/artificial-intelligence/openai-closes-66-billion-funding-haul-valuation-157-billion-with-investment-2024-10-02/ |url-status=live }} It also acquired the domain Chat.com,{{Cite web |last=Wiggers |first=Kyle |date=2024-11-06 |title=OpenAI acquired Chat.com |url=https://techcrunch.com/2024/11/06/openai-acquired-chat-com/ |access-date=2024-12-15 |website=TechCrunch |language=en-US |archive-date=November 27, 2024 |archive-url=https://web.archive.org/web/20241127232242/https://techcrunch.com/2024/11/06/openai-acquired-chat-com/ |url-status=live }}{{cite web |date=2024-11-06 |title=Did OpenAI just spend more than $10 million on a URL? |url=https://www.theverge.com/2024/11/6/24289768/openai-chat-chatgpt-sam-altman-hubspot |access-date=2024-11-10 |publisher=The Verge |archive-date=November 11, 2024 |archive-url=https://web.archive.org/web/20241111205857/https://www.theverge.com/2024/11/6/24289768/openai-chat-chatgpt-sam-altman-hubspot |url-status=live }} and saw the return of Greg Brockman after his brief absence.{{cite news|title=OpenAI co-founder Greg Brockman returns to ChatGPT maker |date=12 November 2024 |url=https://www.reuters.com/technology/artificial-intelligence/openai-co-founder-greg-brockman-returns-ai-startup-bloomberg-news-reports-2024-11-12/ |website=Reuters |access-date=12 November 2024}}

In December 2024, during the "12 Days of OpenAI" event, the company launched the Sora model for ChatGPT Plus and Pro users,{{Cite news |date=2024-12-10 |title=OpenAI releases text-to-video model Sora for ChatGPT Plus and Pro users |url=https://economictimes.indiatimes.com/tech/artificial-intelligence/openai-releases-text-to-video-model-sora-for-chatgpt-plus-and-pro-users/articleshow/116154796.cms |access-date=2024-12-11 |work=The Economic Times |issn=0013-0389 |archive-date=December 11, 2024 |archive-url=https://web.archive.org/web/20241211233708/https://economictimes.indiatimes.com/tech/artificial-intelligence/openai-releases-text-to-video-model-sora-for-chatgpt-plus-and-pro-users/articleshow/116154796.cms |url-status=live }}{{Cite news |date=December 10, 2024 |title=OpenAI releases text-to-video model Sora for ChatGPT Plus and Pro users |url=https://www.reuters.com/technology/artificial-intelligence/openai-releases-text-to-video-model-sora-chatgpt-plus-pro-users-2024-12-09/ |website=Reuters}} It also launched the advanced OpenAI o1 reasoning model{{Cite web |date=2024-12-09 |title=Deep Dive: OpenAI's o1 - The Dawn of Deliberate AI |url=https://portkey.ai/blog/openai-o1-model-card-analysis/ |access-date=2024-12-11 |website=Portkey Blog |language=en |archive-date=December 9, 2024 |archive-url=https://web.archive.org/web/20241209060419/https://portkey.ai/blog/openai-o1-model-card-analysis/ |url-status=live }}{{Cite web |last=Franzen |first=Carl |date=2024-12-05 |title=OpenAI launches full o1 model with image uploads and analysis, debuts ChatGPT Pro |url=https://venturebeat.com/ai/openai-launches-full-o1-model-with-34-reduced-error-rate-debuts-chatgpt-pro/ |access-date=2024-12-11 |website=VentureBeat |language=en-US |archive-date=December 7, 2024 |archive-url=https://web.archive.org/web/20241207181403/https://venturebeat.com/ai/openai-launches-full-o1-model-with-34-reduced-error-rate-debuts-chatgpt-pro/ |url-status=live }} Additionally, ChatGPT Pro—a $200/month subscription service offering unlimited o1 access and enhanced voice features—was introduced, and preliminary benchmark results for the upcoming OpenAI o3 models were shared.{{Cite web |last1=Franzen |first1=Carl |last2=David |first2=Emilia |date=2024-12-20 |title=OpenAI confirms new frontier models o3 and o3-mini |url=https://venturebeat.com/ai/openai-confirms-new-frontier-models-o3-and-o3-mini/ |access-date=2025-01-19 |website=VentureBeat |language=en-US |archive-date=January 20, 2025 |archive-url=https://web.archive.org/web/20250120044522/https://venturebeat.com/ai/openai-confirms-new-frontier-models-o3-and-o3-mini/ |url-status=live }}

On January 20, 2025, DeepSeek released the "DeepSeek-R1" model, which rivaled the performance of OpenAI's o1 and was open-weight.{{Cite web |last=Orland |first=Kyle |date=2025-01-28 |title=How does DeepSeek R1 really fare against OpenAI’s best reasoning models? |url=https://arstechnica.com/ai/2025/01/how-does-deepseek-r1-really-fare-against-openais-best-reasoning-models/ |access-date=2025-02-18 |website=Ars Technica |language=en-US}} DeepSeek claimed that this model only took $5.6 million to train. This news led to panic from investors and caused Nvidia to record the biggest single day market cap loss in history losing $589 billion on January 27.{{Cite news |date=January 27, 2025 |title=Nvidia stock plummets, loses record $589 billion as DeepSeek prompts questions over AI spending |url=https://finance.yahoo.com/news/nvidia-stock-plummets-loses-record-589-billion-as-deepseek-prompts-questions-over-ai-spending-135105824.html |work=Yahoo! finance}}

On January 21, 2025, it was announced that OpenAI, Oracle, SoftBank and MGX would launch The Stargate Project, a joint venture to build an AI infrastructure system in conjunction with the US government. The project takes its name from OpenAI's existing "Stargate" supercomputer project and is estimated to cost $500 billion. The project will be funded over the next four years.{{Cite web |last=Jacobs |first=Jennifer |date=2025-01-22 |title=Trump announces up to $500 billion in private sector AI infrastructure investment - CBS News |url=https://www.cbsnews.com/news/trump-announces-private-sector-ai-infrastructure-investment/ |access-date=2025-01-22 |website=www.cbsnews.com |language=en-US}}

On January 23, OpenAI released Operator, an AI agent and web automation tool for accessing websites to execute goals defined by users. The feature was only available to Pro users in the United States.{{Cite web |last=Metz |first=Cade |date=2025-01-23 |title=OpenAI Unveils A.I. Agent That Can Use Websites on Its Own |url=https://www.nytimes.com/2025/01/23/technology/openai-operator-launch.html |access-date=2025-01-24 |website=New York Times |language=en-US}}{{Cite web |last=Edwards |first=Benj |date=2025-01-23 |title=OpenAI launches Operator, an AI agent that can operate your computer |url=https://arstechnica.com/ai/2025/01/openai-launches-operator-an-ai-agent-that-can-operate-your-computer/ |access-date=2025-01-24 |website=Ars Technica |language=en-US}}

On February 2, OpenAI made a deep research agent, that achieved an accuracy of 26.6 percent on Humanity's Last Exam (HLE) benchmark, available to $200-monthly-fee paying users with up to 100 queries per month, while more “limited access” was promised for Plus, Team and later Enterprise users.

In February, OpenAI underwent a rebranding with a new typeface, word mark, symbol and palette.{{Cite web |last=Bell |first=Jonathan |date=2025-02-04 |title=OpenAI has undergone its first ever rebrand, giving fresh life to ChatGPT interactions |url=https://www.wallpaper.com/tech/openai-has-undergone-its-first-ever-rebrand-giving-fresh-life-to-chatgpt-interactions |access-date=2025-02-05 |website=wallpaper.com |language=en}} OpenAI began collaborating with Broadcom in 2024 to design a custom AI chip capable of both training and inference targeted for mass production in 2026 and to be manufactured by TSMC in 3 nm node. This initiative is intended to reduce OpenAI's dependence on Nvidia GPUs, which are costly and face high demand in the market.

{{cite web |url=https://www.reuters.com/technology/openai-set-finalize-first-custom-chip-design-this-year-2025-02-10/ |title=Exclusive: OpenAI set to finalize first custom chip design this year |date=February 10, 2025 |website=reuters.com |access-date=February 10, 2025 }}

On February 13, Sam Altman announced that GPT-4.5, internally known as "Orion", will be the last model without full chain-of-thought reasoning. Altman also indicated that GPT-5, expected to be released within months, could unify the O-Series and GPT-Series models, eliminating the need to choose between them and phasing out O-series models.{{Cite web |last=Barr |first=Kyle |date=2025-02-20 |title=OpenAI’s GPT-4.5 May Arrive Next Week, but GPT-5 Is Just Around the Corner |url=https://gizmodo.com/openais-gpt-4-5-may-arrive-next-week-but-gpt-5-is-just-around-the-corner-2000566442 |access-date=2025-02-22 |website=Gizmodo |language=en-US}}{{Cite web |last=Nolan |first=Beatrice |title=Sam Altman lays out plans for GPT-5 and GPT-4.5 promising end of 'hated' model picker |url=https://fortune.com/2025/02/14/sam-altman-openai-plans-gpt-5-release-timelines/ |access-date=2025-02-22 |website=Fortune |language=en}}

In March 2025, OpenAI signed an $11.9 billion agreement with CoreWeave, an Nvidia-backed, AI-focused cloud service provider. As part of the deal, OpenAI will receive $350 million worth of CoreWeave shares and gain access to its AI infrastructure, which includes over a quarter million NVIDIA GPUs.{{Cite web |last=Bort |first=Julie |date=2025-03-10 |title=In another chess move with Microsoft, OpenAI is pouring $12B into CoreWeave |url=https://techcrunch.com/2025/03/10/in-another-chess-move-with-microsoft-openai-is-pouring-12b-into-coreweave/ |access-date=2025-03-11 |website=TechCrunch |language=en-US}}

In April 2025, OpenAI raised $40 billion at a $300 billion post-money valuation, marking the largest private technology deal on record. The financing round was led by SoftBank, with other participants including Microsoft, Coatue, Altimeter, and Thrive.{{Cite web |last=Rooney |first=Hayden Field,Kate |date=2025-03-31 |title=OpenAI closes $40 billion funding round, largest private tech deal on record |url=https://www.cnbc.com/2025/03/31/openai-closes-40-billion-in-funding-the-largest-private-fundraise-in-history-softbank-chatgpt.html |access-date=2025-04-01 |website=CNBC |language=en}}{{Cite web |last=Wiggers |first=Kyle |date=2025-03-31 |title=OpenAI raises $40B at $300B post-money valuation |url=https://techcrunch.com/2025/03/31/openai-raises-40b-at-300b-post-money-valuation/ |access-date=2025-04-01 |website=TechCrunch |language=en-US}}

On April 9, 2025, OpenAI countersued Musk in federal court, alleging that he had engaged in "bad-faith tactics" to slow the company’s progress and seize its innovations for his personal benefit. OpenAI also argued that Musk had previously supported the creation of a for-profit structure and had expressed interest in controlling OpenAI himself. The countersuit seeks damages and legal measures to prevent further alleged interference.{{Cite web |last=Habeshian |first=Sareen |date=2025-04-10 |title=OpenAI countersues Elon Musk in bitter legal battle |url=https://www.axios.com/2025/04/10/openai-elon-musk-countersuit |access-date=2025-04-10 |website=Axios |language=en}}

In May 2025, it was reported that OpenAI had agreed to acquire "Windsurf", an AI-assisted coding tool formerly known as "Codeium", for approximately $3 billion.{{cite web |last1=Roof |first1=Katie |last2=Metz |first2=Rachel |title=OpenAI Reaches Agreement to Buy Startup Windsurf for $3 Billion |url=https://www.bloomberg.com/news/articles/2025-05-06/openai-reaches-agreement-to-buy-startup-windsurf-for-3-billion |date=2025-05-06 |website=Bloomberg |access-date=2025-05-20 |language=en}} Windsurf was valued at $1.25 billion in 2024, after a $150 million funding round led by the venture capital firm General Catalyst.{{cite web |last=Metz |first=Cade |title=OpenAI in Talks to Acquire Windsurf, a Programming Tool, for $3 Billion |url=https://www.nytimes.com/2025/05/13/technology/openai-windsurf-talks.html |date=2025-05-13 |website=The New York Times |access-date=2025-05-20 |language=en}}

On May 11, 2025, Financial Times reported that OpenAI and Microsoft are rewriting terms of their multibillion-dollar partnership in a negotiation designed to allow the ChatGPT maker to launch a future IPO, while protecting the software giant's access to cutting-edge AI models.{{Cite news |last=Abraham |first=Rhea Rose |date=2025-05-15 |title=OpenAI negotiates with Microsoft for new funding and future IPO, FT reports |url=https://www.reuters.com/business/openai-negotiates-with-microsoft-unlock-new-funding-future-ipo-ft-reports-2025-05-11/ |access-date=2025-05-15 |work=Reuters}}

File:OpenAI corporate structure.svg

• CEO and co-founder: Sam Altman, former president of the startup accelerator Y Combinator
• President and co-founder: Greg Brockman, former CTO, 3rd employee of Stripe
• Chief Scientist Officer: Jakub Pachocki, former Director of Research at OpenAI
• Chief Operating Officer: Brad Lightcap, previously at Y Combinator and JPMorgan Chase{{Cite news |last=Bordoloi |first=Pritam |date=May 9, 2022 |title=OpenAI gets a new president, CTO & COO in the latest rejig |url=https://analyticsindiamag.com/openai-gets-a-new-president-cto-coo-in-the-latest-rejig// |url-status=live |archive-url=https://web.archive.org/web/20221016010242/https://analyticsindiamag.com/openai-gets-a-new-president-cto-coo-in-the-latest-rejig/ |archive-date=October 16, 2022 |access-date=October 11, 2022 |work=AIM}}
• Chief Financial Officer: Sarah Friar, former Nextdoor CEO and former CFO at Block, Inc.{{Cite news |url=https://www.reuters.com/technology/openai-hires-sarah-friar-cfo-2024-06-10/ |title=OpenAI hires former Nextdoor CEO Sarah Friar as first CFO |date=June 10, 2024 |publisher=Reuters |access-date=June 11, 2024 |archive-date=June 11, 2024 |archive-url=https://web.archive.org/web/20240611124940/https://www.reuters.com/technology/openai-hires-sarah-friar-cfo-2024-06-10/ |url-status=live }}
• Chief Product Officer: Kevin Weil, previously at Twitter, Inc. and Meta Platforms
• Chief Research Officer: Mark Chen, former SVP of Research at OpenAI{{Cite web |last=Levy |first=Hayden Field,Ari |date=2025-03-24 |title=OpenAI expands COO Brad Lightcap's job to include business oversight, as Altman focuses on research |url=https://www.cnbc.com/2025/03/24/openai-expands-coo-brad-lightcaps-job-to-include-business-oversight-.html |access-date=2025-04-28 |website=CNBC |language=en}}
• Chief Compliance Officer: Scott Schools, former Chief Compliance Officer of Uber

{{Anchor|Board}}

• Bret Taylor (chairman), former chairman of Twitter's board of directors and co-CEO of Salesforce
• Sam Altman
• Lawrence Summers, former U.S. Secretary of the Treasury and President of Harvard University
• Adam D'Angelo, co-founder and CEO of Quora
• Sue Desmond-Hellmann, former CEO of the Bill & Melinda Gates Foundation
• Nicole Seligman, attorney and former executive vice president of the Sony Corporation
• Fidji Simo, CEO and chair of Instacart
• Paul Nakasone, former Director of the National Security Agency (2018–2024){{Cite web |last=Peters |first=Jay |date=June 13, 2024 |title=Former head of NSA joins OpenAI board |url=https://www.theverge.com/2024/6/13/24178079/openai-board-paul-nakasone-nsa-safety |url-status=live |archive-url=https://web.archive.org/web/20240614060521/https://www.theverge.com/2024/6/13/24178079/openai-board-paul-nakasone-nsa-safety |archive-date=June 14, 2024 |access-date=June 14, 2024 |website=The Verge |language=en}}
• Zico Kolter, computer scientist{{Cite news |date=August 8, 2024 |title=OpenAI Names Computer Scientist Zico Kolter as New Board Member |url=https://www.bloomberg.com/news/articles/2024-08-08/openai-names-computer-scientist-zico-kolter-as-new-board-member |access-date=August 14, 2024 |work=Bloomberg.com |language=en |archive-date=August 8, 2024 |archive-url=https://archive.today/20240808194030/https://www.bloomberg.com/news/articles/2024-08-08/openai-names-computer-scientist-zico-kolter-as-new-board-member |url-status=live }}
• Adebayo Ogunlesi, managing partner at Global Infrastructure Partners{{Cite news |last1=Criddle |first1=Cristina |last2=Gara |first2=Antoine |date=2025-01-14 |title=OpenAI appoints one of Wall Street's most powerful dealmakers to its board |url=https://www.ft.com/content/63b08a9d-e537-4d60-9904-c59958a16982 |access-date=2025-01-15 |work=Financial Times}}

Sources:{{Cite news |date=March 11, 2024 |title=Who are OpenAI's new board members? |url=https://www.reuters.com/technology/who-are-openais-new-board-members-2024-03-11/ |work=Reuters |archive-date=March 12, 2024 |access-date=August 29, 2024 |archive-url=https://archive.today/20240312110610/https://www.reuters.com/technology/who-are-openais-new-board-members-2024-03-11/ |url-status=live }}

Source:

• Reid Hoffman, LinkedIn co-founder{{cite news|last1=Liedtke|first1=Michael|title=Elon Musk, Peter Thiel, Reid Hoffman, others back $1 billion OpenAI research center|url=http://www.mercurynews.com/business/ci_29256196/elon-musk-peter-thiel-reid-hoffman-others-back|access-date=December 19, 2015|work=Mercury News|archive-date=December 22, 2015|archive-url=https://web.archive.org/web/20151222152100/http://www.mercurynews.com/business/ci_29256196/elon-musk-peter-thiel-reid-hoffman-others-back |url-status=live}}
• Peter Thiel, PayPal co-founder
• Jessica Livingston, a founding partner of Y Combinator
• Elon Musk, co-founder

Some scientists, such as Stephen Hawking and Stuart Russell, have articulated concerns that if advanced AI gains the ability to redesign itself at an ever-increasing rate, an unstoppable "intelligence explosion" could lead to human extinction. Co-founder Musk characterizes AI as humanity's "biggest existential threat".{{cite news |last1=Piper |first1=Kelsey |title=Why Elon Musk fears artificial intelligence |url=https://www.vox.com/future-perfect/2018/11/2/18053418/elon-musk-artificial-intelligence-google-deepmind-openai |access-date=March 10, 2021 |work=Vox |date=November 2, 2018 |language=en |archive-date=April 23, 2021 |archive-url=https://web.archive.org/web/20210423095445/https://www.vox.com/future-perfect/2018/11/2/18053418/elon-musk-artificial-intelligence-google-deepmind-openai |url-status=live }}

Musk and Altman have stated they are partly motivated by concerns about AI safety and the existential risk from artificial general intelligence.{{cite news|last1=Lewontin|first1=Max|title=Open AI: Effort to democratize artificial intelligence research?|url=http://www.csmonitor.com/Technology/2015/1214/Open-AI-Effort-to-democratize-artificial-intelligence-research|access-date=December 19, 2015|work=The Christian Science Monitor |date=December 14, 2015|archive-date=December 19, 2015|archive-url=https://web.archive.org/web/20151219071134/http://www.csmonitor.com/Technology/2015/1214/Open-AI-Effort-to-democratize-artificial-intelligence-research|url-status=live}} OpenAI states that "it's hard to fathom how much human-level AI could benefit society", and that it is equally difficult to comprehend "how much it could damage society if built or used incorrectly". Research on safety cannot safely be postponed: "because of AI's surprising history, it's hard to predict when human-level AI might come within reach."{{cite news|last1=Mendoza|first1=Jessica|title=Tech leaders launch nonprofit to save the world from killer robots|url=http://www.csmonitor.com/Science/2015/1214/Tech-leaders-launch-nonprofit-to-save-the-world-from-killer-robots|work=The Christian Science Monitor|access-date=December 19, 2015|archive-date=July 3, 2018|archive-url=https://web.archive.org/web/20180703170556/https://www.csmonitor.com/Science/2015/1214/Tech-leaders-launch-nonprofit-to-save-the-world-from-killer-robots|url-status=live}} OpenAI states that AI "should be an extension of individual human wills and, in the spirit of liberty, as broadly and evenly distributed as possible." Co-chair Sam Altman expects the decades-long project to surpass human intelligence.{{cite magazine|last1=Metz|first1=Cade |title=Elon Musk's Billion-Dollar AI Plan Is About Far More Than Saving the World|url=https://www.wired.com/2015/12/elon-musks-billion-dollar-ai-plan-is-about-far-more-than-saving-the-world/ |access-date=December 19, 2015|magazine=Wired |date=December 15, 2015|quote=Altman said they expect this decades-long project to surpass human intelligence. |archive-date=December 19, 2015|archive-url=https://web.archive.org/web/20151219155928/http://www.wired.com/2015/12/elon-musks-billion-dollar-ai-plan-is-about-far-more-than-saving-the-world/|url-status=live}}

Vishal Sikka, former CEO of Infosys, stated that an "openness", where the endeavor would "produce results generally in the greater interest of humanity", was a fundamental requirement for his support; and that OpenAI "aligns very nicely with our long-held values" and their "endeavor to do purposeful work".{{cite web|author1=Vishal Sikka|title=OpenAI: AI for All |website=InfyTalk|publisher=Infosys|date=December 14, 2015 |url=http://www.infosysblogs.com/infytalk/2015/12/openai_ai_for_all.html |access-date=December 22, 2015 |url-status=dead |archive-url=https://web.archive.org/web/20151222094518/http://www.infosysblogs.com/infytalk/2015/12/openai_ai_for_all.html|archive-date=December 22, 2015|author1-link=Vishal Sikka}} Cade Metz of Wired suggested that corporations such as Amazon might be motivated by a desire to use open-source software and data to level the playing field against corporations such as Google and Facebook, which own enormous supplies of proprietary data. Altman stated that Y Combinator companies would share their data with OpenAI.

In the early years before his 2018 departure, Musk posed the question: "What is the best thing we can do to ensure the future is good? We could sit on the sidelines or we can encourage regulatory oversight, or we could participate with the right structure with people who care deeply about developing AI in a way that is safe and is beneficial to humanity." He acknowledged that "there is always some risk that in actually trying to advance (friendly) AI we may create the thing we are concerned about"; but nonetheless, that the best defense was "to empower as many people as possible to have AI. If everyone has AI powers, then there's not any one person or a small set of individuals who can have AI superpower."{{cite news |title=Silicon Valley investors to bankroll artificial-intelligence center |date=December 13, 2015 |url=http://www.seattletimes.com/business/technology/silicon-valley-investors-to-bankroll-artificial-intelligence-center/ |access-date=December 19, 2015 |work=The Seattle Times |archive-date=January 5, 2016|archive-url=https://web.archive.org/web/20160105181552/http://www.seattletimes.com/business/technology/silicon-valley-investors-to-bankroll-artificial-intelligence-center/ |url-status=live}}

Musk and Altman's counterintuitive strategy—that of trying to reduce the potential harm of AI by giving everyone access to it—is controversial among those concerned with existential risk from AI. Philosopher Nick Bostrom said, "If you have a button that could do bad things to the world, you don't want to give it to everyone."{{cite magazine|author1=Cade Metz|title=Inside OpenAI, Elon Musk's Wild Plan to Set Artificial Intelligence Free|url=https://www.wired.com/2016/04/openai-elon-musk-sam-altman-plan-to-set-artificial-intelligence-free/ |access-date=April 28, 2016|magazine=Wired|date=April 27, 2016|language=en-US|archive-date=April 27, 2016|archive-url=https://web.archive.org/web/20160427162700/http://www.wired.com/2016/04/openai-elon-musk-sam-altman-plan-to-set-artificial-intelligence-free/ |url-status=live}} During a 2016 conversation about technological singularity, Altman said, "We don't plan to release all of our source code" and mentioned a plan to "allow wide swaths of the world to elect representatives to a new governance board". Greg Brockman stated, "Our goal right now{{nbsp}}... is to do the best thing there is to do. It's a little vague."{{cite magazine|title=Sam Altman's Manifest Destiny |url=http://www.newyorker.com/magazine/2016/10/10/sam-altmans-manifest-destiny|access-date=October 4, 2016|magazine=The New Yorker|issue=October 10, 2016|url-status=live|archive-url=https://web.archive.org/web/20161004091200/http://www.newyorker.com/magazine/2016/10/10/sam-altmans-manifest-destiny|archive-date=October 4, 2016}}

Conversely, OpenAI's initial decision to withhold GPT-2 around 2019, due to a wish to "err on the side of caution" in the presence of potential misuse, was criticized by advocates of openness. Delip Rao, an expert in text generation, stated, "I don't think {{bracket|OpenAI}} spent enough time proving {{bracket|GPT-2}} was actually dangerous." Other critics argued that open publication was necessary to replicate the research and to create countermeasures.{{cite news |last1=Vincent |first1=James |title=AI researchers debate the ethics of sharing potentially harmful programs |url=https://www.theverge.com/2019/2/21/18234500/ai-ethics-debate-researchers-harmful-programs-openai |access-date=March 6, 2020 |work=The Verge |date=February 21, 2019 |language=en |archive-date=February 9, 2021 |archive-url=https://web.archive.org/web/20210209123243/https://www.theverge.com/2019/2/21/18234500/ai-ethics-debate-researchers-harmful-programs-openai |url-status=live }}

More recently, in 2022, OpenAI published its approach to the alignment problem, anticipating that aligning AGI to human values would likely be harder than aligning current AI systems: "Unaligned AGI could pose substantial risks to humanity[,] and solving the AGI alignment problem could be so difficult that it will require all of humanity to work together". They stated that they intended to explore how to better use human feedback to train AI systems, and how to safely use AI to incrementally automate alignment research.{{Cite web |title=Our approach to alignment research |url=https://openai.com/blog/our-approach-to-alignment-research |access-date=April 26, 2023 |website=openai.com |language=en-US |archive-date=April 26, 2023 |archive-url=https://web.archive.org/web/20230426111006/https://openai.com/blog/our-approach-to-alignment-research |url-status=live }}

In 2024, following the temporary removal of Sam Altman and his return, many employees gradually left OpenAI, including most of the original leadership team and a significant number of AI safety researchers.{{Cite web |last=Piper |first=Kelsey |date=2024-10-28 |title=Inside OpenAI's multibillion-dollar gambit to become a for-profit company |url=https://www.vox.com/future-perfect/380117/openai-microsoft-sam-altman-nonprofit-for-profit-foundation-artificial-intelligence |access-date=2024-12-03 |website=Vox |language=en-US |archive-date=December 3, 2024 |archive-url=https://web.archive.org/web/20241203052435/https://www.vox.com/future-perfect/380117/openai-microsoft-sam-altman-nonprofit-for-profit-foundation-artificial-intelligence |url-status=live }}{{Cite magazine |last=Knight |first=Will |title=The OpenAI Talent Exodus Gives Rivals an Opening |url=https://www.wired.com/story/openai-departures-research-rivals-artificial-intelligence/ |access-date=2024-12-03 |magazine=Wired |language=en-US |issn=1059-1028 |archive-date=December 4, 2024 |archive-url=https://web.archive.org/web/20241204012149/https://www.wired.com/story/openai-departures-research-rivals-artificial-intelligence/ |url-status=live }} OpenAI also planned a restructuring to operate as a for-profit company. This restructuring could grant Altman a stake in the company.{{Cite web |last=Edwards |first=Benj |date=2024-09-26 |title=OpenAI plans tectonic shift from nonprofit to for-profit, giving Altman equity |url=https://arstechnica.com/information-technology/2024/09/openai-plans-tectonic-shift-from-nonprofit-to-for-profit-giving-altman-equity/ |access-date=2024-12-03 |website=Ars Technica |language=en-US |archive-date=December 2, 2024 |archive-url=https://web.archive.org/web/20241202204712/https://arstechnica.com/information-technology/2024/09/openai-plans-tectonic-shift-from-nonprofit-to-for-profit-giving-altman-equity/ |url-status=live }}

In February 2025, OpenAI CEO Sam Altman stated that the company is interested in collaborating with the People's Republic of China, despite regulatory restrictions imposed by the U.S. government.{{cite news |date=February 2025 |title=OpenAI keen to work with China, CEO Sam Altman says, as DeepSeek rattles the tech market |url=https://www.scmp.com/tech/big-tech/article/3298396/openai-keen-work-china-ceo-sam-altman-says-deepseek-rattles-tech-market |access-date=February 17, 2025 |work=South China Morning Post}} This shift comes in response to the growing influence of the Chinese artificial intelligence company DeepSeek, which has disrupted the AI market with advanced models, including DeepSeek V3 and DeepSeek R1, known for their efficiency and cost-effectiveness.{{cite news |date=February 2025 |title=DeepSeek spurs Baidu, other AI competitors to adopt open-source strategy |url=https://www.scmp.com/tech/tech-trends/article/3298739/deepseek-spurs-baidu-other-ai-competitors-adopt-open-source-strategy |access-date=February 17, 2025 |work=South China Morning Post}}

The emergence of DeepSeek has led major Chinese tech firms such as Baidu and others to embrace an open-source strategy, intensifying competition with OpenAI. Altman acknowledged the uncertainty regarding U.S. government approval for AI cooperation with China but emphasized the importance of fostering dialogue between technological leaders in both nations.{{cite news |date=February 2025 |title=OpenAI CEO discusses AI collaboration and regulatory challenges |url=https://www.scmp.com/tech/big-tech/article/3298396/openai-keen-work-china-ceo-sam-altman-says-deepseek-rattles-tech-market |access-date=February 17, 2025 |work=South China Morning Post}}

At its beginning, OpenAI's research included many projects focused on reinforcement learning (RL).{{Cite web |last=Wiggers |first=Kyle |date=July 16, 2021 |title=OpenAI disbands its robotics research team |url=https://venturebeat.com/business/openai-disbands-its-robotics-research-team/ |url-status=live |archive-url=https://web.archive.org/web/20230212150930/https://venturebeat.com/business/openai-disbands-its-robotics-research-team/ |archive-date=February 12, 2023 |access-date=February 12, 2023 |website=VentureBeat |language=en-US}} OpenAI has been viewed as an important competitor to DeepMind.{{cite news |last1=Lee |first1=Dave |date=October 15, 2019 |title=Robot solves Rubik's cube, but not grand challenge |url=https://www.bbc.com/news/technology-50064225 |url-status=live |archive-url=https://web.archive.org/web/20200403202722/https://www.bbc.com/news/technology-50064225 |archive-date=April 3, 2020 |access-date=February 29, 2020 |work=BBC News}}

Announced in 2016, Gym was an open-source Python library designed to facilitate the development of reinforcement learning algorithms. It aimed to standardize how environments are defined in AI research, making published research more easily reproducible{{cite news|author1=Dave Gershgorn|title=Elon Musk's Artificial Intelligence Group Opens A 'Gym' To Train A.I. |url=http://www.popsci.com/elon-musks-artificial-intelligence-group-opens-gym-to-train-ai|access-date=April 29, 2016|work=Popular Science |date=April 27, 2016|archive-date=April 30, 2016|archive-url=https://web.archive.org/web/20160430045139/http://www.popsci.com/elon-musks-artificial-intelligence-group-opens-gym-to-train-ai|url-status=live}}{{cite web|author1=Greg Brockman|author2=John Schulman |title=OpenAI Gym Beta|url=https://openai.com/blog/openai-gym-beta/|website=OpenAI Blog|publisher=OpenAI|access-date=April 29, 2016|language=en-us|date=April 27, 2016|archive-date=February 26, 2019|archive-url=https://web.archive.org/web/20190226173517/https://openai.com/blog/openai-gym-beta/|url-status=live}} while providing users with a simple interface for interacting with these environments. In 2022, new developments of Gym have been moved to the library Gymnasium.{{Cite web |title=openai/gym |website=GitHub |url=https://github.com/openai/gym |access-date=August 29, 2024 |quote=The team that has been maintaining Gym since 2021 has moved all future development to Gymnasium, a drop in replacement for Gym (import gymnasium as gym), and Gym will not be receiving any future updates. |archive-date=August 23, 2024 |archive-url=https://web.archive.org/web/20240823033813/https://github.com/openai/gym |url-status=live }}{{Cite web |date=October 25, 2022 |title=Announcing The Farama Foundation - The future of open source reinforcement learning |url=https://farama.org/Announcing-The-Farama-Foundation |access-date=August 29, 2024 |website=The Farama Foundation |language=en |archive-date=August 29, 2024 |archive-url=https://web.archive.org/web/20240829061404/https://farama.org/Announcing-The-Farama-Foundation |url-status=live }}

Released in 2018, Gym Retro is a platform for reinforcement learning (RL) research on video games{{Cite web |date=May 25, 2018 |title=Gym Retro |url=https://openai.com/blog/gym-retro/ |url-status=live |archive-url=https://web.archive.org/web/20230212155510/https://openai.com/blog/gym-retro/ |archive-date=February 12, 2023 |access-date=February 12, 2023 |website=OpenAI |language=en}} using RL algorithms and study generalization. Prior RL research focused mainly on optimizing agents to solve single tasks. Gym Retro gives the ability to generalize between games with similar concepts but different appearances.

Released in 2017, RoboSumo is a virtual world where humanoid metalearning robot agents initially lack knowledge of how to even walk, but are given the goals of learning to move and to push the opposing agent out of the ring.{{cite magazine |date=October 11, 2017 |title=AI Sumo Wrestlers Could Make Future Robots More Nimble |url=https://www.wired.com/story/ai-sumo-wrestlers-could-make-future-robots-more-nimble/ |url-status=live |archive-url=https://web.archive.org/web/20171107024652/https://www.wired.com/story/ai-sumo-wrestlers-could-make-future-robots-more-nimble/ |archive-date=November 7, 2017 |access-date=November 2, 2017 |magazine=Wired}} Through this adversarial learning process, the agents learn how to adapt to changing conditions. When an agent is then removed from this virtual environment and placed in a new virtual environment with high winds, the agent braces to remain upright, suggesting it had learned how to balance in a generalized way.{{cite news |title=OpenAI's Goofy Sumo-Wrestling Bots Are Smarter Than They Look |url=https://www.technologyreview.com/the-download/609117/openais-goofy-sumo-wrestling-bots-are-smarter-than-they-look/ |url-status=live |archive-url=https://web.archive.org/web/20181109021143/https://www.technologyreview.com/the-download/609117/openais-goofy-sumo-wrestling-bots-are-smarter-than-they-look/ |archive-date=November 9, 2018 |access-date=November 2, 2017 |work=MIT Technology Review |language=en}} OpenAI's Igor Mordatch argued that competition between agents could create an intelligence "arms race" that could increase an agent's ability to function even outside the context of the competition.

{{Main|OpenAI Five}}

OpenAI Five is a team of five OpenAI-curated bots used in the competitive five-on-five video game Dota 2, that learn to play against human players at a high skill level entirely through trial-and-error algorithms. Before becoming a team of five, the first public demonstration occurred at The International 2017, the annual premiere championship tournament for the game, where Dendi, a professional Ukrainian player, lost against a bot in a live one-on-one matchup.{{cite web |last1=Savov |first1=Vlad |date=August 14, 2017 |title=My favorite game has been invaded by killer AI bots and Elon Musk hype |url=https://www.theverge.com/2017/8/14/16141938/dota-2-openai-bots-elon-musk-artificial-intelligence |url-status=live |archive-url=https://web.archive.org/web/20180626030145/https://www.theverge.com/2017/8/14/16141938/dota-2-openai-bots-elon-musk-artificial-intelligence |archive-date=June 26, 2018 |access-date=June 25, 2018 |website=The Verge}}{{cite web |last1=Frank |first1=Blair Hanley |title=OpenAI's bot beats top Dota 2 player so badly that he quits |url=https://venturebeat.com/2017/08/11/openais-bot-beats-top-dota-2-player-so-badly-that-he-quits/ |url-status=dead |archive-url=https://web.archive.org/web/20170812065202/https://venturebeat.com/2017/08/11/openais-bot-beats-top-dota-2-player-so-badly-that-he-quits/ |archive-date=August 12, 2017 |access-date=August 12, 2017 |website=Venture Beat}} After the match, CTO Greg Brockman explained that the bot had learned by playing against itself for two weeks of real time, and that the learning software was a step in the direction of creating software that can handle complex tasks like a surgeon.{{cite web |date=August 11, 2017 |title=Dota 2 |url=https://blog.openai.com/dota-2/ |url-status=live |archive-url=https://web.archive.org/web/20170811235617/https://blog.openai.com/dota-2/ |archive-date=August 11, 2017 |access-date=August 12, 2017 |website=blog.openai.com}}{{cite web |date=August 16, 2017 |title=More on Dota 2 |url=https://blog.openai.com/more-on-dota-2/ |url-status=live |archive-url=https://web.archive.org/web/20190223171009/https://blog.openai.com/more-on-dota-2/ |archive-date=February 23, 2019 |access-date=August 16, 2017 |website=blog.openai.com}} The system uses a form of reinforcement learning, as the bots learn over time by playing against themselves hundreds of times a day for months, and are rewarded for actions such as killing an enemy and taking map objectives.{{cite magazine |last1=Simonite |first1=Tom |title=Can Bots Outwit Humans in One of the Biggest Esports Games? |url=https://www.wired.com/story/can-bots-outwit-humans-in-one-of-the-biggest-esports-games/ |url-status=live |magazine=Wired |archive-url=https://web.archive.org/web/20180625213810/https://www.wired.com/story/can-bots-outwit-humans-in-one-of-the-biggest-esports-games/ |archive-date=June 25, 2018 |access-date=June 25, 2018}}{{cite news |last1=Kahn |first1=Jeremy |date=June 25, 2018 |title=A Bot Backed by Elon Musk Has Made an AI Breakthrough in Video Game World |newspaper=Bloomberg.com |publisher=Bloomberg L.P. |url=https://www.bloomberg.com/news/articles/2018-06-25/musk-backed-bot-conquers-e-gamer-teams-in-ai-breakthrough |url-status=live |access-date=June 27, 2018 |archive-url=https://web.archive.org/web/20180627144300/https://www.bloomberg.com/news/articles/2018-06-25/musk-backed-bot-conquers-e-gamer-teams-in-ai-breakthrough |archive-date=June 27, 2018}}{{cite web |last1=Clifford |first1=Catherine |date=June 28, 2018 |title=Bill Gates says gamer bots from Elon Musk-backed nonprofit are 'huge milestone' in A.I. |url=https://www.cnbc.com/2018/06/27/bill-gates-openai-robots-beating-humans-at-dota-2-is-ai-milestone.html |url-status=live |archive-url=https://web.archive.org/web/20180628231125/https://www.cnbc.com/2018/06/27/bill-gates-openai-robots-beating-humans-at-dota-2-is-ai-milestone.html |archive-date=June 28, 2018 |access-date=June 29, 2018 |publisher=CNBC}}

By June 2018, the ability of the bots expanded to play together as a full team of five, and they were able to defeat teams of amateur and semi-professional players.{{cite web |date=July 18, 2018 |title=OpenAI Five Benchmark |url=https://blog.openai.com/openai-five-benchmark/ |url-status=live |archive-url=https://web.archive.org/web/20190213165905/https://blog.openai.com/openai-five-benchmark/ |archive-date=February 13, 2019 |access-date=August 25, 2018 |website=blog.openai.com}}{{cite web |last1=Vincent |first1=James |date=June 25, 2018 |title=AI bots trained for 180 years a day to beat humans at Dota 2 |url=https://www.theverge.com/2018/6/25/17492918/openai-dota-2-bot-ai-five-5v5-matches |url-status=live |archive-url=https://web.archive.org/web/20180625183203/https://www.theverge.com/2018/6/25/17492918/openai-dota-2-bot-ai-five-5v5-matches |archive-date=June 25, 2018 |access-date=June 25, 2018 |website=The Verge}}{{cite web |last1=Savov |first1=Vlad |date=August 6, 2018 |title=The OpenAI Dota 2 bots just defeated a team of former pros |url=https://www.theverge.com/2018/8/6/17655086/dota2-openai-bots-professional-gaming-ai |url-status=live |archive-url=https://web.archive.org/web/20180807113227/https://www.theverge.com/2018/8/6/17655086/dota2-openai-bots-professional-gaming-ai |archive-date=August 7, 2018 |access-date=August 7, 2018 |website=The Verge}} At The International 2018, OpenAI Five played in two exhibition matches against professional players, but ended up losing both games.{{cite magazine |last1=Simonite |first1=Tom |title=Pro Gamers Fend off Elon Musk-Backed AI Bots—for Now |url=https://www.wired.com/story/pro-gamers-fend-off-elon-musks-ai-bots/ |url-status=live |magazine=Wired |archive-url=https://web.archive.org/web/20180824120523/https://www.wired.com/story/pro-gamers-fend-off-elon-musks-ai-bots/ |archive-date=August 24, 2018 |access-date=August 25, 2018}}{{cite web |last1=Quach |first1=Katyanna |title=Game over, machines: Humans defeat OpenAI bots once again at video games Olympics |url=https://www.theregister.co.uk/2018/08/24/openai_bots_eliminated_dota_2/ |url-status=live |archive-url=https://web.archive.org/web/20180825110329/https://www.theregister.co.uk/2018/08/24/openai_bots_eliminated_dota_2/ |archive-date=August 25, 2018 |access-date=August 25, 2018 |website=The Register}}{{cite web |date=August 24, 2018 |title=The International 2018: Results |url=https://blog.openai.com/the-international-2018-results/ |url-status=live |archive-url=https://web.archive.org/web/20180824131639/https://blog.openai.com/the-international-2018-results/ |archive-date=August 24, 2018 |access-date=August 25, 2018 |website=blog.openai.com}} In April 2019, OpenAI Five defeated OG, the reigning world champions of the game at the time, 2:0 in a live exhibition match in San Francisco.{{cite web |last1=Statt |first1=Nick |date=April 13, 2019 |title=OpenAI's Dota 2 AI steamrolls world champion e-sports team with back-to-back victories |url=https://www.theverge.com/2019/4/13/18309459/openai-five-dota-2-finals-ai-bot-competition-og-e-sports-the-international-champion |url-status=live |archive-url=https://web.archive.org/web/20190415011925/https://www.theverge.com/2019/4/13/18309459/openai-five-dota-2-finals-ai-bot-competition-og-e-sports-the-international-champion |archive-date=April 15, 2019 |access-date=July 20, 2019 |website=The Verge}}{{cite web |date=April 15, 2019 |title=How to Train Your OpenAI Five |url=https://openai.com/blog/how-to-train-your-openai-five/ |url-status=live |archive-url=https://web.archive.org/web/20190630013455/https://openai.com/blog/how-to-train-your-openai-five/ |archive-date=June 30, 2019 |access-date=July 20, 2019 |website=OpenAI Blog}} The bots' final public appearance came later that month, where they played in 42,729 total games in a four-day open online competition, winning 99.4% of those games.{{cite web |last1=Wiggers |first1=Kyle |date=April 22, 2019 |title=OpenAI's Dota 2 bot defeated 99.4% of players in public matches. |url=https://venturebeat.com/2019/04/22/openais-dota-2-bot-defeated-99-4-of-players-in-public-matches/ |url-status=live |archive-url=https://web.archive.org/web/20190711151127/https://venturebeat.com/2019/04/22/openais-dota-2-bot-defeated-99-4-of-players-in-public-matches/ |archive-date=July 11, 2019 |access-date=April 22, 2019 |website=Venture Beat}}

OpenAI Five's mechanisms in Dota 2's bot player shows the challenges of AI systems in multiplayer online battle arena (MOBA) games and how OpenAI Five has demonstrated the use of deep reinforcement learning (DRL) agents to achieve superhuman competence in Dota 2 matches.{{Cite book |last1=Fangasadha |first1=Edbert Felix |last2=Soeroredjo |first2=Steffi |last3=Anderies |last4=Gunawan |first4=Alexander Agung Santoso |title=2022 International Seminar on Application for Technology of Information and Communication (ISemantic) |chapter=Literature Review of OpenAI Five's Mechanisms in Dota 2's Bot Player |date=September 17, 2022 |url=https://ieeexplore.ieee.org/document/9920480 |publisher=IEEE |pages=183–190 |doi=10.1109/iSemantic55962.2022.9920480 |isbn=978-1-6654-8837-2 |s2cid=253047170 |access-date=August 7, 2023 |archive-date=May 12, 2023 |archive-url=https://web.archive.org/web/20230512224139/https://ieeexplore.ieee.org/document/9920480 |url-status=live }}

Developed in 2018, Dactyl uses machine learning to train a Shadow Hand, a human-like robot hand, to manipulate physical objects.{{Cite web |last=Vincent |first=James |date=July 30, 2018 |title=OpenAI sets new benchmark for robot dexterity |url=https://www.theverge.com/2018/7/30/17621112/openai-robot-dexterity-dactyl-artificial-intelligence |access-date=February 12, 2023 |website=The Verge |language=en-US |archive-date=February 12, 2023 |archive-url=https://web.archive.org/web/20230212152342/https://www.theverge.com/2018/7/30/17621112/openai-robot-dexterity-dactyl-artificial-intelligence |url-status=live }} It learns entirely in simulation using the same RL algorithms and training code as OpenAI Five. OpenAI tackled the object orientation problem by using domain randomization, a simulation approach which exposes the learner to a variety of experiences rather than trying to fit to reality. The set-up for Dactyl, aside from having motion tracking cameras, also has RGB cameras to allow the robot to manipulate an arbitrary object by seeing it. In 2018, OpenAI showed that the system was able to manipulate a cube and an octagonal prism.{{cite arXiv |author=OpenAI |last2=Andrychowicz |first2=Marcin |last3=Baker |first3=Bowen |last4=Chociej |first4=Maciej |last5=Józefowicz |first5=Rafał |last6=McGrew |first6=Bob |last7=Pachocki |first7=Jakub |last8=Petron |first8=Arthur |last9=Plappert |first9=Matthias |last10=Powell |first10=Glenn |last11=Ray |first11=Alex |last12=Schneider |first12=Jonas |last13=Sidor |first13=Szymon |last14=Tobin |first14=Josh |last15=Welinder |first15=Peter |last16=Weng |first16=Lilian |last17=Zaremba |first17=Wojciech |year=2019 |title=Learning Dexterous In-Hand Manipulation |eprint=1808.00177v5 |class=cs.LG}}

In 2019, OpenAI demonstrated that Dactyl could solve a Rubik's Cube. The robot was able to solve the puzzle 60% of the time. Objects like the Rubik's Cube introduce complex physics that is harder to model. OpenAI did this by improving the robustness of Dactyl to perturbations by using Automatic Domain Randomization (ADR), a simulation approach of generating progressively more difficult environments. ADR differs from manual domain randomization by not needing a human to specify randomization ranges.{{cite arXiv |author1=OpenAI |last2=Akkaya |first2=Ilge |last3=Andrychowicz |first3=Marcin |last4=Chociej |first4=Maciek |last5=Litwin |first5=Mateusz |last6=McGrew |first6=Bob |last7=Petron |first7=Arthur |last8=Paino |first8=Alex |last9=Plappert |first9=Matthias |last10=Powell |first10=Glenn |last11=Ribas |first11=Raphael |year=2019 |title=Solving Rubik's Cube with a Robot Hand |eprint=1910.07113v1 |class=cs.LG}}

In June 2020, OpenAI announced a multi-purpose API which it said was "for accessing new AI models developed by OpenAI" to let developers call on it for "any English language AI task".{{Cite web |date=June 11, 2020 |url=https://techcrunch.com/2020/06/11/openai-makes-an-all-purpose-api-for-its-text-based-ai-capabilities/ |url-status=live |archive-url=https://web.archive.org/web/20200612041855/https://techcrunch.com/2020/06/11/openai-makes-an-all-purpose-api-for-its-text-based-ai-capabilities/ |archive-date=June 12, 2020 |access-date=June 11, 2020 |work=TechCrunch |quote=If you’ve ever wanted to try out OpenAI's vaunted machine learning toolset, it just got a lot easier. The company has released an API that lets developers call its AI tools in on "virtually any English language task."|title=OpenAI makes an all-purpose API for its text-based AI capabilities|author-first1=Devin|author-last1=Coldewey}}

The company has popularized generative pretrained transformers (GPT).{{cite web | url=https://www.makeuseof.com/gpt-models-explained-and-compared/ | title=GPT-1 to GPT-4: Each of OpenAI's GPT Models Explained and Compared | date=April 11, 2023 | access-date=April 29, 2023 | archive-date=April 15, 2023 | archive-url=https://web.archive.org/web/20230415175013/https://www.makeuseof.com/gpt-models-explained-and-compared/ | url-status=live }}

{{further|Generative pre-trained transformer#History}}

File:Full GPT architecture.svg

The original paper on generative pre-training of a transformer-based language model was written by Alec Radford and his colleagues, and published in preprint on OpenAI's website on June 11, 2018.{{Cite web |title=Improving Language Understanding by Generative Pre-Training |url=https://cdn.openai.com/research-covers/language-unsupervised/language_understanding_paper.pdf |access-date=June 9, 2020|archive-date=January 26, 2021|archive-url=https://web.archive.org/web/20210126024542/https://cdn.openai.com/research-covers/language-unsupervised/language_understanding_paper.pdf|url-status=live}} It showed how a generative model of language could acquire world knowledge and process long-range dependencies by pre-training on a diverse corpus with long stretches of contiguous text.

{{main|GPT-2}}

File:GPT2-talks-about-GPT2.png

Generative Pre-trained Transformer 2 ("GPT-2") is an unsupervised transformer language model and the successor to OpenAI's original GPT model ("GPT-1"). GPT-2 was announced in February 2019, with only limited demonstrative versions initially released to the public. The full version of GPT-2 was not immediately released due to concern about potential misuse, including applications for writing fake news.{{Cite news| last=Hern| first=Alex| title=New AI fake text generator may be too dangerous to release, say creators| newspaper=The Guardian| date=February 14, 2019| url=https://www.theguardian.com/technology/2019/feb/14/elon-musk-backed-ai-writes-convincing-news-fiction| access-date=February 14, 2019| archive-date=February 14, 2019| archive-url=https://web.archive.org/web/20190214173112/https://www.theguardian.com/technology/2019/feb/14/elon-musk-backed-ai-writes-convincing-news-fiction| url-status=live}} Some experts expressed skepticism that GPT-2 posed a significant threat.

In response to GPT-2, the Allen Institute for Artificial Intelligence responded with a tool to detect "neural fake news".{{cite news |last1=Schwartz |first1=Oscar |title=Could 'fake text' be the next global political threat? |url=https://www.theguardian.com/technology/2019/jul/04/ai-fake-text-gpt-2-concerns-false-information |access-date=July 16, 2019 |work=The Guardian |date=July 4, 2019 |archive-date=July 16, 2019 |archive-url=https://web.archive.org/web/20190716035703/https://www.theguardian.com/technology/2019/jul/04/ai-fake-text-gpt-2-concerns-false-information |url-status=live }} Other researchers, such as Jeremy Howard, warned of "the technology to totally fill Twitter, email, and the web up with reasonable-sounding, context-appropriate prose, which would drown out all other speech and be impossible to filter".{{cite news |last1=Vincent |first1=James |title=OpenAI's new multitalented AI writes, translates, and slanders |url=https://www.theverge.com/2019/2/14/18224704/ai-machine-learning-language-models-read-write-openai-gpt2 |access-date=July 16, 2019 |work=The Verge |date=February 14, 2019 |archive-date=December 18, 2020 |archive-url=https://web.archive.org/web/20201218091707/https://www.theverge.com/2019/2/14/18224704/ai-machine-learning-language-models-read-write-openai-gpt2 |url-status=live }} In November 2019, OpenAI released the complete version of the GPT-2 language model.{{Cite web |url=https://openai.com/blog/gpt-2-1-5b-release/ |title=GPT-2: 1.5B Release|date=November 5, 2019|website=OpenAI|language=en|access-date=November 14, 2019|archive-date=November 14, 2019|archive-url=https://web.archive.org/web/20191114074358/https://openai.com/blog/gpt-2-1-5b-release/|url-status=live}} Several websites host interactive demonstrations of different instances of GPT-2 and other transformer models.{{Cite web |url=https://transformer.huggingface.co/|title=Write With Transformer |access-date=December 4, 2019|archive-date=December 4, 2019|archive-url=https://web.archive.org/web/20191204060111/https://transformer.huggingface.co/|url-status=live}}{{Cite web |url=https://talktotransformer.com/|title=Talk to Transformer|access-date=December 4, 2019|archive-date=December 4, 2019|archive-url=https://web.archive.org/web/20191204015009/https://talktotransformer.com/|url-status=live}}{{Cite web |url=https://creativeengines.ai/|title=CreativeEngines|access-date=June 25, 2021|archive-date=February 3, 2023|archive-url=https://web.archive.org/web/20230203201104/https://creativeengines.ai/|url-status=live}}

GPT-2's authors argue unsupervised language models to be general-purpose learners, illustrated by GPT-2 achieving state-of-the-art accuracy and perplexity on 7 of 8 zero-shot tasks (i.e. the model was not further trained on any task-specific input-output examples).

The corpus it was trained on, called WebText, contains slightly 40 gigabytes of text from URLs shared in Reddit submissions with at least 3 upvotes. It avoids certain issues encoding vocabulary with word tokens by using byte pair encoding. This permits representing any string of characters by encoding both individual characters and multiple-character tokens.{{citation |title=Language Models are Unsupervised Multitask Learners |url=https://d4mucfpksywv.cloudfront.net/better-language-models/language_models_are_unsupervised_multitask_learners.pdf |access-date=December 4, 2019 |archive-date=December 12, 2019 |archive-url=https://web.archive.org/web/20191212223916/https://d4mucfpksywv.cloudfront.net/better-language-models/language_models_are_unsupervised_multitask_learners.pdf |url-status=live }}

{{main|GPT-3}}

First described in May 2020, Generative Pre-trained{{efn|The term "pre-training" refers to general language training as distinct from fine-tuning for specific tasks.}} Transformer 3 (GPT-3) is an unsupervised transformer language model and the successor to GPT-2.{{Cite web |date=May 29, 2020 |title=openai/gpt-3 |url=https://github.com/openai/gpt-3 |url-status=live |archive-url=https://web.archive.org/web/20201114165742/https://github.com/openai/gpt-3 |archive-date=November 14, 2020 |access-date=May 29, 2020 |publisher=OpenAI}}{{Cite web |last=Sagar |first=Ram |date=June 3, 2020 |title=OpenAI Releases GPT-3, The Largest Model So Far |url=https://analyticsindiamag.com/open-ai-gpt-3-language-model/ |url-status=live |archive-url=https://web.archive.org/web/20200804173452/https://analyticsindiamag.com/open-ai-gpt-3-language-model/ |archive-date=August 4, 2020 |access-date=June 14, 2020 |website=Analytics India Magazine |language=en-US}} OpenAI stated that the full version of GPT-3 contained 175 billion parameters, two orders of magnitude larger than the 1.5 billion{{citation |title=Language Models are Unsupervised Multitask Learners |quote=GPT-2, is a 1.5B parameter Transformer |url=https://d4mucfpksywv.cloudfront.net/better-language-models/language_models_are_unsupervised_multitask_learners.pdf |access-date=December 4, 2019 |archive-date=December 12, 2019 |archive-url=https://web.archive.org/web/20191212223916/https://d4mucfpksywv.cloudfront.net/better-language-models/language_models_are_unsupervised_multitask_learners.pdf |url-status=live }} in the full version of GPT-2 (although GPT-3 models with as few as 125 million parameters were also trained).{{Cite arXiv|last1=Brown|first1=Tom |last2=Mann |first2=Benjamin |last3=Ryder |first3=Nick |last4=Subbiah |first4=Melanie |last5=Kaplan |first5=Jared |last6=Dhariwal |first6=Prafulla |last7=Neelakantan |first7=Arvind |last8=Shyam |first8=Pranav |last9=Sastry |first9=Girish |last10=Askell |first10=Amanda |last11=Agarwal |first11=Sandhini |date=June 1, 2020 |title=Language Models are Few-Shot Learners |class=cs.CL |eprint=2005.14165 |quote=Since we increase the capacity by over two orders of magnitude from GPT-2 to GPT-3}}

OpenAI stated that GPT-3 succeeded at certain "meta-learning" tasks and could generalize the purpose of a single input-output pair. The GPT-3 release paper gave examples of translation and cross-linguistic transfer learning between English and Romanian, and between English and German.{{Cite arXiv|last1=Brown|first1=Tom |last2=Mann |first2=Benjamin |last3=Ryder |first3=Nick |last4=Subbiah |first4=Melanie |last5=Kaplan |first5=Jared |last6=Dhariwal |first6=Prafulla |last7=Neelakantan |first7=Arvind |last8=Shyam |first8=Pranav |last9=Sastry |first9=Girish |last10=Askell |first10=Amanda |last11=Agarwal |first11=Sandhini |date=June 1, 2020 |title=Language Models are Few-Shot Learners |eprint=2005.14165 |page=appendix |class=cs.CL}}

GPT-3 dramatically improved benchmark results over GPT-2. OpenAI cautioned that such scaling-up of language models could be approaching or encountering the fundamental capability limitations of predictive language models.{{cite news |last1=Ray |first1=Tiernan |title=OpenAI's gigantic GPT-3 hints at the limits of language models for AI |url=https://www.zdnet.com/article/openais-gigantic-gpt-3-hints-at-the-limits-of-language-models-for-ai/ |access-date=June 5, 2020 |publisher=ZDNet |year=2020 |language=en |archive-date=June 1, 2020 |archive-url=https://web.archive.org/web/20200601081629/https://www.zdnet.com/article/openais-gigantic-gpt-3-hints-at-the-limits-of-language-models-for-ai/ |url-status=live }} Pre-training GPT-3 required several thousand petaflop/s-days{{efn|One petaflop/s-day is approximately equal to 10²⁰ neural net operations.{{cite web |title=AI and Compute |date=May 16, 2018 |last1=Amodei |first1=Dario |last2=Hernandez |first2=Danny |quote=A petaflop/s-day (pfs-day) consists of performing 10¹⁵ neural net operations per second for one day, or a total of about 10²⁰ operations. The compute-time product serves as a mental convenience, similar to kW-hr for energy. |url=https://openai.com/blog/ai-and-compute/#fn2 |access-date=August 30, 2020 |archive-date=June 17, 2020 |archive-url=https://web.archive.org/web/20200617200602/https://openai.com/blog/ai-and-compute/#fn2 |url-status=live }}}} of compute, compared to tens of petaflop/s-days for the full GPT-2 model. Like its predecessor, the GPT-3 trained model was not immediately released to the public for concerns of possible abuse, although OpenAI planned to allow access through a paid cloud API after a two-month free private beta that began in June 2020.{{Cite web |date=June 11, 2020|title=OpenAI API|url=https://openai.com/blog/openai-api/|access-date=June 14, 2020|website=OpenAI|language=en|quote=Why did OpenAI choose to release an API instead of open-sourcing the models?
There are three main reasons we did this. First, commercializing the technology helps us pay for our ongoing AI research, safety, and policy efforts. Second, many of the models underlying the API are very large, taking a lot of expertise to develop and deploy and making them very expensive to run. This makes it hard for anyone except larger companies to benefit from the underlying technology. We're hopeful that the API will make powerful AI systems more accessible to smaller businesses and organizations. Third, the API model allows us to more easily respond to misuse of the technology. Since it is hard to predict the downstream use cases of our models, it feels inherently safer to release them via an API and broaden access over time, rather than release an open source model where access cannot be adjusted if it turns out to have harmful applications.|archive-date=June 11, 2020|archive-url=https://web.archive.org/web/20200611150951/https://openai.com/blog/openai-api/|url-status=live}}{{Cite web |last=Eadicicco |first=Lisa |title=The artificial intelligence company that Elon Musk helped found is now selling the text-generation software it previously said was too dangerous to launch|url=https://www.businessinsider.com/elon-musk-openai-sell-text-tool-it-said-was-dangerous-2020-6 |access-date=July 6, 2020 |website=Business Insider|archive-date=November 14, 2020|archive-url=https://web.archive.org/web/20201114205255/https://www.businessinsider.com/elon-musk-openai-sell-text-tool-it-said-was-dangerous-2020-6|url-status=live}}

On September 23, 2020, GPT-3 was licensed exclusively to Microsoft.{{Cite web|title=OpenAI is giving Microsoft exclusive access to its GPT-3 language model |url=https://www.technologyreview.com/2020/09/23/1008729/openai-is-giving-microsoft-exclusive-access-to-its-gpt-3-language-model/ |access-date=September 24, 2020|website=MIT Technology Review |language=en|archive-date=February 5, 2021|archive-url=https://web.archive.org/web/20210205121656/https://www.technologyreview.com/2020/09/23/1008729/openai-is-giving-microsoft-exclusive-access-to-its-gpt-3-language-model/|url-status=live}}{{Cite web|date=September 22, 2020|title=Microsoft gets exclusive license for OpenAI's GPT-3 language model|url=https://venturebeat.com/2020/09/22/microsoft-gets-exclusive-license-for-openais-gpt-3-language-model/|access-date=September 24, 2020|website=VentureBeat|language=en-US|archive-date=November 8, 2020|archive-url=https://web.archive.org/web/20201108090524/https://venturebeat.com/2020/09/22/microsoft-gets-exclusive-license-for-openais-gpt-3-language-model/|url-status=live}}

{{main article|OpenAI Codex}}

Announced in mid-2021, Codex is a descendant of GPT-3 that has additionally been trained on code from 54 million GitHub repositories,{{Cite news |last=Alford |first=Anthony |date=August 31, 2021 |title=OpenAI Announces 12 Billion Parameter Code-Generation AI Codex |url=https://www.infoq.com/news/2021/08/openai-codex/ |url-status=live |archive-url=https://web.archive.org/web/20220709221205/https://www.infoq.com/news/2021/08/openai-codex/ |archive-date=July 9, 2022 |access-date=September 3, 2021 |work=InfoQ}}{{Cite news |last=Wiggers |first=Kyle |date=July 8, 2021 |title=OpenAI warns AI behind GitHub's Copilot may be susceptible to bias |url=https://venturebeat.com/2021/07/08/openai-warns-ai-behind-githubs-copilot-may-be-susceptible-to-bias/ |url-status=live |archive-url=https://web.archive.org/web/20230203201912/https://venturebeat.com/business/openai-warns-ai-behind-githubs-copilot-may-be-susceptible-to-bias/ |archive-date=February 3, 2023 |access-date=September 3, 2021 |work=VentureBeat}} and is the AI powering the code autocompletion tool GitHub Copilot. In August 2021, an API was released in private beta.{{cite web |last=Zaremba |first=Wojciech |author-link=Wojciech Zaremba |date=August 10, 2021 |title=OpenAI Codex |url=https://openai.com/blog/openai-codex/ |url-status=live |archive-url=https://web.archive.org/web/20230203201912/https://openai.com/blog/openai-codex/ |archive-date=February 3, 2023 |access-date=September 3, 2021 |website=OpenAI}} According to OpenAI, the model can create working code in over a dozen programming languages, most effectively in Python.

Several issues with glitches, design flaws and security vulnerabilities were cited.{{Cite news |last=Dickson |first=Ben |date=August 16, 2021 |title=What to expect from OpenAI's Codex API |url=https://venturebeat.com/2021/08/16/what-to-expect-from-openais-codex-api/ |url-status=live |archive-url=https://web.archive.org/web/20230203201913/https://venturebeat.com/ai/what-to-expect-from-openais-codex-api/ |archive-date=February 3, 2023 |access-date=September 3, 2021 |work=VentureBeat}}{{Cite news |last=Claburn |first=Thomas |date=August 25, 2021 |title=GitHub's Copilot may steer you into dangerous waters about 40% of the time – study |url=https://www.theregister.com/2021/08/25/github_copilot_study/ |url-status=live |archive-url=https://web.archive.org/web/20230203201913/https://www.theregister.com/2021/08/25/github_copilot_study/ |archive-date=February 3, 2023 |access-date=September 3, 2021 |work=The Register}}

OpenAI announced that they would discontinue support for Codex API on March 23, 2023.{{cite web |date=March 21, 2023 |title=OpenAI Might Invite Legal Trouble |url=https://analyticsindiamag.com/openai-might-invite-legal-trouble/ |url-status=live |archive-url=https://web.archive.org/web/20230323205407/https://analyticsindiamag.com/openai-might-invite-legal-trouble/ |archive-date=March 23, 2023 |access-date=March 23, 2023 |website=Analytics India Magazine}}

{{main|GPT-4}}

On March 14, 2023, OpenAI announced the release of Generative Pre-trained Transformer 4 (GPT-4), capable of accepting text or image inputs.{{Cite web |last=Vincent |first=James |date=March 14, 2023 |title=OpenAI announces GPT-4—the next generation of its AI language model |url=https://www.theverge.com/2023/3/14/23638033/openai-gpt-4-chatgpt-multimodal-deep-learning |url-status=live |archive-url=https://web.archive.org/web/20230314195326/https://www.theverge.com/2023/3/14/23638033/openai-gpt-4-chatgpt-multimodal-deep-learning |archive-date=March 14, 2023 |access-date=March 14, 2023 |website=The Verge |language=en-US}} They announced that the updated technology passed a simulated law school bar exam with a score around the top 10% of test takers. (By contrast, GPT-3.5 scored around the bottom 10%.) They said that GPT-4 could also read, analyze or generate up to 25,000 words of text, and write code in all major programming languages.{{Cite web |last=Wiggers |first=Kyle |date=March 14, 2023 |title=OpenAI releases GPT-4, a multimodal AI that it claims is state-of-the-art |url=https://techcrunch.com/2023/03/14/openai-releases-gpt-4-ai-that-it-claims-is-state-of-the-art/ |url-status=live |archive-url=https://web.archive.org/web/20230315003723/https://techcrunch.com/2023/03/14/openai-releases-gpt-4-ai-that-it-claims-is-state-of-the-art/ |archive-date=March 15, 2023 |access-date=March 14, 2023 |website=TechCrunch |language=en-US}}

Observers reported that the iteration of ChatGPT using GPT-4 was an improvement on the previous GPT-3.5-based iteration, with the caveat that GPT-4 retained some of the problems with earlier revisions.{{cite news |last1=Belfield |first1=Haydn |date=March 25, 2023 |title=If your AI model is going to sell, it has to be safe |url=https://www.vox.com/future-perfect/2023/3/25/23655082/ai-openai-gpt-4-safety-microsoft-facebook-meta |url-status=live |archive-url=https://web.archive.org/web/20230328192017/https://www.vox.com/future-perfect/2023/3/25/23655082/ai-openai-gpt-4-safety-microsoft-facebook-meta |archive-date=March 28, 2023 |access-date=March 30, 2023 |work=Vox}} GPT-4 is also capable of taking images as input on ChatGPT.{{Cite news |last=Roose |first=Kevin |date=September 28, 2023 |title=The New ChatGPT Can 'See' and 'Talk.' Here's What It's Like. |url=https://www.nytimes.com/2023/09/27/technology/new-chatgpt-can-see-hear.html |url-status=live |archive-url=https://web.archive.org/web/20231031055345/https://www.nytimes.com/2023/09/27/technology/new-chatgpt-can-see-hear.html |archive-date=October 31, 2023 |access-date=December 1, 2023 |work=The New York Times}} OpenAI has declined to reveal various technical details and statistics about GPT-4, such as the precise size of the model.{{Cite web |last=Vincent |first=James |date=March 15, 2023 |title=OpenAI co-founder on company's past approach to openly sharing research: "We were wrong" |url=https://www.theverge.com/2023/3/15/23640180/openai-gpt-4-launch-closed-research-ilya-sutskever-interview |url-status=live |archive-url=https://web.archive.org/web/20230317210900/https://www.theverge.com/2023/3/15/23640180/openai-gpt-4-launch-closed-research-ilya-sutskever-interview |archive-date=March 17, 2023 |access-date=March 18, 2023 |website=The Verge}}

On May 13, 2024, OpenAI announced and released GPT-4o, which can process and generate text, images and audio.{{Cite web |last=Wiggers |first=Kyle |date=May 13, 2024 |title=OpenAI debuts GPT-4o 'omni' model now powering ChatGPT |url=https://techcrunch.com/2024/05/13/openais-newest-model-is-gpt-4o/ |access-date=May 13, 2024 |website=TechCrunch |language=en-US |archive-date=May 22, 2024 |archive-url=https://web.archive.org/web/20240522094111/https://techcrunch.com/2024/05/13/openais-newest-model-is-gpt-4o/ |url-status=live }} GPT-4o achieved state-of-the-art results in voice, multilingual, and vision benchmarks, setting new records in audio speech recognition and translation.{{cite web |last1=van Rijmenam |first1=Mark |date=May 13, 2024 |title=OpenAI Launched GPT-4o: The Future of AI Interactions Is Here |url=https://www.thedigitalspeaker.com/openai-gpt4o-future-ai-interactions/ |access-date=May 17, 2024 |website=The Digital Speaker |archive-date=May 17, 2024 |archive-url=https://web.archive.org/web/20240517013424/https://www.thedigitalspeaker.com/openai-gpt4o-future-ai-interactions/ |url-status=live }}{{Cite web |last=Daws |first=Ryan |date=May 14, 2024 |title=GPT-4o delivers human-like AI interaction with text, audio, and vision integration |url=https://www.artificialintelligence-news.com/2024/05/14/gpt-4o-human-like-ai-interaction-text-audio-vision-integration/ |access-date=May 18, 2024 |website=AI News |language=en-GB |archive-date=May 18, 2024 |archive-url=https://web.archive.org/web/20240518131148/https://www.artificialintelligence-news.com/2024/05/14/gpt-4o-human-like-ai-interaction-text-audio-vision-integration/ |url-status=live }} It scored 88.7% on the Massive Multitask Language Understanding (MMLU) benchmark compared to 86.5% by GPT-4.{{Cite web |title=Hello GPT-4o |url=https://openai.com/index/hello-gpt-4o/ |website=OpenAI |access-date=July 14, 2024 |archive-date=May 14, 2024 |archive-url=https://web.archive.org/web/20240514024319/https://openai.com/index/hello-gpt-4o/ |url-status=live }}

On July 18, 2024, OpenAI released GPT-4o mini, a smaller version of GPT-4o replacing GPT-3.5 Turbo on the ChatGPT interface. Its API costs $0.15 per million input tokens and $0.60 per million output tokens, compared to $5 and $15 respectively for GPT-4o. OpenAI expects it to be particularly useful for enterprises, startups and developers seeking to automate services with AI agents.{{Cite web |last=Franzen |first=Carl |date=July 18, 2024 |title=OpenAI unveils GPT-4o mini — a smaller, much cheaper multimodal AI model |url=https://venturebeat.com/ai/openai-unveils-gpt-4o-mini-a-smaller-much-cheaper-multimodal-ai-model/ |access-date=July 18, 2024 |website=VentureBeat |language=en-US |archive-date=July 18, 2024 |archive-url=https://web.archive.org/web/20240718185315/https://venturebeat.com/ai/openai-unveils-gpt-4o-mini-a-smaller-much-cheaper-multimodal-ai-model/ |url-status=live }}

In March 2025, OpenAI released GPT-4o's native image generation feature, as an alternative to DALL-E 3. {{Cite web |last=Zeff |first=Kyle Wiggers, Maxwell |date=2025-03-25 |title=ChatGPT's image-generation feature gets an upgrade |url=https://techcrunch.com/2025/03/25/chatgpts-image-generation-feature-gets-an-upgrade/ |access-date=2025-03-27 |website=TechCrunch |language=en-US}}

On February 27, 2025, OpenAI released GPT-4.5, codenamed Orion. Sam Altman claimed that GPT-4.5 would present inaccurate information less frequently than previous models, and described it as a "giant, expensive model".{{Cite web |last=Novet |first=Jordan |date=2025-02-27 |title=OpenAI launching GPT-4.5, its next general-purpose large language model |url=https://www.cnbc.com/2025/02/27/openai-launching-gpt-4point5-general-purpose-large-language-model.html |access-date=2025-03-18 |website=CNBC |language=en}}

On April 14, 2025, OpenAI released the GPT-4.1 model. They also released two “smaller, faster, and cheaper” models including GPT-4.1 mini and GPT-4.1 nano.{{Cite web |last=Weatherbed |first=Jess |date=2025-04-14 |title=OpenAI debuts its GPT-4.1 flagship AI model |url=https://www.theverge.com/news/647896/openai-chatgpt-gpt-4-1-mini-nano-launch-availability |access-date=2025-04-15 |website=The Verge |language=en-US}}{{Cite web |title=Introducing GPT-4.1 in the API |url=https://openai.com/index/gpt-4-1/ |access-date=2025-04-15 |website=openai.com |language=en-US}}{{Cite news |last=Knight |first=Will |title=OpenAI’s New GPT 4.1 Models Excel at Coding |url=https://www.wired.com/story/openai-announces-4-1-ai-model-coding/ |access-date=2025-04-15 |work=Wired |language=en-US |issn=1059-1028}}

{{Main article|OpenAI o1}}

On September 12, 2024, OpenAI released the o1-preview and o1-mini models, which have been designed to take more time to think about their responses, leading to higher accuracy. These models are particularly effective in science, coding, and reasoning tasks, and were made available to ChatGPT Plus and Team members.{{Cite magazine |last=Knight |first=Will |title=OpenAI Announces a New AI Model, Code-Named Strawberry, That Solves Difficult Problems Step by Step |url=https://www.wired.com/story/openai-o1-strawberry-problem-reasoning/ |access-date=September 14, 2024 |magazine=Wired |language=en-US |issn=1059-1028 |archive-date=September 14, 2024 |archive-url=https://web.archive.org/web/20240914005013/https://www.wired.com/story/openai-o1-strawberry-problem-reasoning/ |url-status=live }}{{Cite web |last=Robison |first=Kylie |date=September 12, 2024 |title=OpenAI releases o1, its first model with 'reasoning' abilities |url=https://www.theverge.com/2024/9/12/24242439/openai-o1-model-reasoning-strawberry-chatgpt |access-date=September 17, 2024 |website=The Verge |language=en |archive-date=September 13, 2024 |archive-url=https://web.archive.org/web/20240913134303/https://www.theverge.com/2024/9/12/24242439/openai-o1-model-reasoning-strawberry-chatgpt |url-status=live }} In December 2024, o1-preview was replaced by o1.{{Cite web |last=Franzen |first=Carl |date=2024-12-05 |title=OpenAI launches full o1 model with image uploads and analysis, debuts ChatGPT Pro |url=https://venturebeat.com/ai/openai-launches-full-o1-model-with-34-reduced-error-rate-debuts-chatgpt-pro/ |access-date=2024-12-07 |website=VentureBeat |language=en-US |archive-date=December 7, 2024 |archive-url=https://web.archive.org/web/20241207181403/https://venturebeat.com/ai/openai-launches-full-o1-model-with-34-reduced-error-rate-debuts-chatgpt-pro/ |url-status=live }} In March 2025, the o1-Pro model was made available through OpenAI's developer API, which was previously available to ChatGPT Pro users since December 2024. The pricing is $150 per million input tokens and $600 per million output tokens.{{Cite web |last=Wiggers |first=Kyle |date=2025-03-19 |title=OpenAI's o1-pro is the company's most expensive AI model yet |url=https://techcrunch.com/2025/03/19/openais-o1-pro-is-its-most-expensive-model-yet/ |access-date=2025-03-21 |website=TechCrunch |language=en-US}}

{{main|OpenAI o3}}

On December 20, 2024, OpenAI unveiled o3, the successor of the o1 reasoning model. OpenAI also unveiled o3-mini, a lighter and faster version of OpenAI o3. As of December 21, 2024, this model is not available for public use. According to OpenAI, they are testing o3 and o3-mini.{{Cite web |last1=Zeff |first1=Maxwell |last2=Wiggers |first2=Kyle |date=2024-12-20 |title=OpenAI announces new o3 models |url=https://techcrunch.com/2024/12/20/openai-announces-new-o3-model/ |access-date=2024-12-20 |website=TechCrunch |language=en-US |archive-date=December 20, 2024 |archive-url=https://web.archive.org/web/20241220201039/https://techcrunch.com/2024/12/20/openai-announces-new-o3-model/ |url-status=live }}{{Cite magazine |last=Knight |first=Will |title=OpenAI Upgrades Its Smartest AI Model With Improved Reasoning Skills |url=https://www.wired.com/story/openai-o3-reasoning-model-google-gemini/ |access-date=2024-12-20 |magazine=Wired |language=en-US |issn=1059-1028 |archive-date=December 20, 2024 |archive-url=https://web.archive.org/web/20241220200335/https://www.wired.com/story/openai-o3-reasoning-model-google-gemini/ |url-status=live }} Until January 10, 2025, safety and security researchers had the opportunity to apply for early access to these models.{{Cite web |date=December 20, 2024 |title=Early access for safety testing |url=https://openai.com/index/early-access-for-safety-testing/ |access-date=2024-12-23 |website=OpenAI |language=en-US |archive-date=December 21, 2024 |archive-url=https://web.archive.org/web/20241221081849/https://openai.com/index/early-access-for-safety-testing/ |url-status=live }} The model is called o3 rather than o2 to avoid confusion with telecommunications services provider O2.{{Cite web |last1=Zeff |first1=Maxwell |last2=Wiggers |first2=Kyle |date=2024-12-20 |title=OpenAI announces new o3 models |url=https://techcrunch.com/2024/12/20/openai-announces-new-o3-model/ |access-date=2024-12-23 |website=TechCrunch |language=en-US |archive-date=December 20, 2024 |archive-url=https://web.archive.org/web/20241220201039/https://techcrunch.com/2024/12/20/openai-announces-new-o3-model/ |url-status=live }} On April 2025, OpenAI released o3 to all the paid users. o3 has enhance reasoning and problem-solving capabilities than o1.{{Cite news |agency = The Hindu Bureau |date=2025-04-17 |title=OpenAI launches new AI reasoning models o3 and 04-mini; older models to be phased out |url=https://www.thehindu.com/sci-tech/technology/openai-launches-new-ai-reasoning-models-o3-and-04-mini-older-models-to-be-phased-out/article69459565.ece |access-date=2025-04-17 |work=The Hindu |language=en-IN |issn=0971-751X}}

{{main|ChatGPT Deep Research}}

Deep research is an agent developed by OpenAI, unveiled on February 2, 2025. It leverages the capabilities of OpenAI's o3 model to perform extensive web browsing, data analysis, and synthesis, delivering comprehensive reports within a timeframe of 5 to 30 minutes.{{Cite web|url=https://www.reuters.com/technology/openai-launches-new-ai-tool-facilitate-research-tasks-2025-02-03/|title=OpenAI launches new AI tool to facilitate research tasks|work=Reuters |date=February 3, 2025|via=www.reuters.com}} With browsing and Python tools enabled, it reached an accuracy of 26.6 percent on HLE (Humanity's Last Exam) benchmark.{{Cite web |last=Lawler |first=Richard |date=2025-02-03 |title=ChatGPT's agent can now do deep research for you |url=https://www.theverge.com/news/604902/chagpt-deep-research-ai-agent |access-date=2025-02-05 |website=The Verge |language=en-US}} In April 2025, OpenAI started rolling out a lightweight version of Deep Research to all its ChatGPT free users.{{Cite web |date=2025-04-25 |title=OpenAI rolls out free, lightweight Deep Research tool for all ChatGPT users |url=https://www.indiatoday.in/technology/news/story/openai-rolls-out-free-lightweight-deep-research-tool-for-all-chatgpt-users-2715040-2025-04-25 |access-date=2025-04-25 |website=India Today |language=en}}{{Cite news |date=2025-04-25 |title=OpenAI introduces cost-efficient, lightweight version of ChatGPT research tool |url=https://timesofindia.indiatimes.com/technology/artificial-intelligence/openai-introduces-cost-efficient-lightweight-version-of-chatgpt-research-tool/articleshow/120603631.cms |access-date=2025-04-25 |work=The Times of India |issn=0971-8257}}

Revealed in 2021, CLIP (Contrastive Language–Image Pre-training) is a model that is trained to analyze the semantic similarity between text and images. It can notably be used for image classification.{{cite web |date=January 5, 2021 |title=CLIP: Connecting Text and Images |url=https://openai.com/blog/clip/ |url-status=live |archive-url=https://web.archive.org/web/20210325202038/https://openai.com/blog/clip/ |archive-date=March 25, 2021 |access-date=March 27, 2021}}

{{main|DALL-E}}

File:DALL-E sample.png

Revealed in 2021, DALL-E is a Transformer model that creates images from textual descriptions.{{cite web|url=https://openai.com/blog/dall-e/|title=DALL·E: Creating Images from Text|date=January 5, 2021|access-date=March 27, 2021|archive-date=March 27, 2021|archive-url=https://web.archive.org/web/20210327133043/https://openai.com/blog/dall-e/|url-status=live}} DALL-E uses a 12-billion-parameter version of GPT-3 to interpret natural language inputs (such as "a green leather purse shaped like a pentagon" or "an isometric view of a sad capybara") and generate corresponding images. It can create images of realistic objects ("a stained-glass window with an image of a blue strawberry") as well as objects that do not exist in reality ("a cube with the texture of a porcupine"). As of March 2021, no API or code is available.

In April 2022, OpenAI announced DALL-E 2, an updated version of the model with more realistic results.{{cite web |title=DALL·E 2 |url=https://openai.com/dall-e-2/ |access-date=April 6, 2022 |website=OpenAI |language=en |archive-date=April 6, 2022 |archive-url=https://web.archive.org/web/20220406141035/https://openai.com/dall-e-2/ |url-status=live }} In December 2022, OpenAI published on GitHub software for Point-E, a new rudimentary system for converting a text description into a 3-dimensional model.{{cite news |title=ChatGPT: A scientist explains the hidden genius and pitfalls of OpenAI's chatbot |url=https://www.sciencefocus.com/news/chatgpt-scientist-openai-chatbot/ |access-date=December 30, 2022 |work=BBC Science Focus Magazine |date=2022 |language=en |archive-date=February 3, 2023 |archive-url=https://web.archive.org/web/20230203201910/https://www.sciencefocus.com/news/chatgpt-scientist-openai-chatbot/ |url-status=live }}

In September 2023, OpenAI announced DALL-E 3, a more powerful model better able to generate images from complex descriptions without manual prompt engineering and render complex details like hands and text.{{cite news |title=OpenAI's new AI image generator pushes the limits in detail and prompt fidelity |url=https://arstechnica.com/information-technology/2023/09/openai-announces-dall-e-3-a-next-gen-ai-image-generator-based-on-chatgpt/ |access-date=November 21, 2023 |archive-url=https://web.archive.org/web/20231116152507/https://arstechnica.com/information-technology/2023/09/openai-announces-dall-e-3-a-next-gen-ai-image-generator-based-on-chatgpt/ |archive-date=November 16, 2023 }} It was released to the public as a ChatGPT Plus feature in October.{{cite web |title=DALL·E 3 is now available in ChatGPT Plus and Enterprise |url=https://openai.com/blog/dall-e-3-is-now-available-in-chatgpt-plus-and-enterprise |access-date=November 21, 2023 |archive-url=https://web.archive.org/web/20231120155249/https://openai.com/blog/dall-e-3-is-now-available-in-chatgpt-plus-and-enterprise |archive-date=November 20, 2023 }}

{{main|Sora (text-to-video model)}}

Sora is a text-to-video model that can generate videos based on short descriptive prompts{{cite news |last=Metz |first=Cade |date=February 15, 2024 |title=OpenAI Unveils A.I. That Instantly Generates Eye-Popping Videos |url=https://www.nytimes.com/2024/02/15/technology/openai-sora-videos.html |url-status=live |archive-url=https://web.archive.org/web/20240215220626/https://www.nytimes.com/2024/02/15/technology/openai-sora-videos.html |archive-date=February 15, 2024 |access-date=February 16, 2024 |work=The New York Times |publisher=}} as well as extend existing videos forwards or backwards in time.{{cite web |date=February 15, 2024 |title=Video generation models as world simulators |url=https://openai.com/research/video-generation-models-as-world-simulators |url-status=live |archive-url=https://web.archive.org/web/20240216072133/https://openai.com/research/video-generation-models-as-world-simulators |archive-date=February 16, 2024 |access-date=February 16, 2024 |publisher=OpenAI}} It can generate videos with resolution up to 1920x1080 or 1080x1920. The maximal length of generated videos is unknown.

Sora's development team named it after the Japanese word for "sky", to signify its "limitless creative potential". Sora's technology is an adaptation of the technology behind the DALL·E 3 text-to-image model.{{cite web |last1=Brooks |first1=Tim |last2=Peebles |first2=Bill |last3=Holmes |first3=Connor |last4=DePue |first4=Will |last5=Guo |first5=Yufei |last6=Jing |first6=Li |last7=Schnurr |first7=David |last8=Taylor |first8=Joe |last9=Luhman |first9=Troy |date=February 15, 2024 |title=Video generation models as world simulators |url=https://openai.com/research/video-generation-models-as-world-simulators |url-status=live |archive-url=https://web.archive.org/web/20240216072133/https://openai.com/research/video-generation-models-as-world-simulators |archive-date=February 16, 2024 |access-date=February 16, 2024 |website=Openai.com |publisher=OpenAI |first10=Eric |last10=Luhman |first11=Clarence Wing Yin |last11=Ng |first12=Ricky |last12=Wang |first13=Aditya |last13=Ramesh}} OpenAI trained the system using publicly-available videos as well as copyrighted videos licensed for that purpose, but did not reveal the number or the exact sources of the videos.

OpenAI demonstrated some Sora-created high-definition videos to the public on February 15, 2024, stating that it could generate videos up to one minute long. It also shared a technical report highlighting the methods used to train the model, and the model's capabilities. It acknowledged some of its shortcomings, including struggles simulating complex physics.{{cite news |last=Pequeño IV |first=Antonio |date=February 15, 2024 |title=OpenAI Reveals 'Sora': AI Video Model Capable Of Realistic Text-To-Video Prompts |url=https://www.forbes.com/sites/antoniopequenoiv/2024/02/15/openai-reveals-sora-ai-video-model-capable-of-realistic-text-to-video-prompts/ |url-status=live |archive-url=https://web.archive.org/web/20240215220634/https://www.forbes.com/sites/antoniopequenoiv/2024/02/15/openai-reveals-sora-ai-video-model-capable-of-realistic-text-to-video-prompts/ |archive-date=February 15, 2024 |access-date=February 16, 2024 |work=Forbes |publisher=}} Will Douglas Heaven of the MIT Technology Review called the demonstration videos "impressive", but noted that they must have been cherry-picked and might not represent Sora's typical output.

Despite skepticism from some academic leaders following Sora's public demo, notable entertainment-industry figures have shown significant interest in the technology's potential. In an interview, actor/filmmaker Tyler Perry expressed his astonishment at the technology's ability to generate realistic video from text descriptions, citing its potential to revolutionize storytelling and content creation. He said that his excitement about Sora's possibilities was so strong that he had decided to pause plans for expanding his Atlanta-based movie studio.{{Cite web |last=Clark |first=Elijah |title=Tyler Perry Warns Of AI Threat After Sora Debut Halts An $800 Million Studio Expansion |url=https://www.forbes.com/sites/elijahclark/2024/02/23/tyler-perry-warns-of-ai-threat-to-jobs-after-viewing-openai-sora/ |access-date=March 24, 2024 |website=Forbes |language=en}}

{{anchor|OpenAI Whisper}}

{{main|Whisper (speech recognition system)}}

Released in 2022, Whisper is a general-purpose speech recognition model.{{Cite web |last=Wiggers |first=Kyle |date=September 21, 2022 |title=OpenAI open-sources Whisper, a multilingual speech recognition system |url=https://techcrunch.com/2022/09/21/openai-open-sources-whisper-a-multilingual-speech-recognition-system/ |url-status=live |archive-url=https://web.archive.org/web/20230212154543/https://techcrunch.com/2022/09/21/openai-open-sources-whisper-a-multilingual-speech-recognition-system/ |archive-date=February 12, 2023 |access-date=February 12, 2023 |website=TechCrunch |language=en-US}} It is trained on a large dataset of diverse audio and is also a multi-task model that can perform multilingual speech recognition as well as speech translation and language identification.{{Cite arXiv |eprint=2212.04356 |class=eess.AS |first1=Alec |last1=Radford |first2=Jong Wook |last2=Kim |title=Robust Speech Recognition via Large-Scale Weak Supervision |last3=Xu |first3=Tao |last4=Brockman |first4=Greg |last5=McLeavey |first5=Christine |last6=Sutskever |first6=Ilya |year=2022 |page=}}

Released in 2019, MuseNet is a deep neural net trained to predict subsequent musical notes in MIDI music files. It can generate songs with 10 instruments in 15 styles. According to The Verge, a song generated by MuseNet tends to start reasonably but then fall into chaos the longer it plays.{{cite news |date=April 2019 |title=OpenAI's MuseNet generates AI music at the push of a button |work=The Verge |url=https://www.theverge.com/2019/4/26/18517803/openai-musenet-artificial-intelligence-ai-music-generation-lady-gaga-harry-potter-mozart |url-status=live |archive-url=https://web.archive.org/web/20190628164236/https://www.theverge.com/2019/4/26/18517803/openai-musenet-artificial-intelligence-ai-music-generation-lady-gaga-harry-potter-mozart |archive-date=June 28, 2019 |access-date=June 8, 2020}}{{cite web |date=April 25, 2019 |title=MuseNet |publisher=OpenAI |url=https://openai.com/blog/musenet/ |url-status=live |archive-url=https://web.archive.org/web/20200613055143/https://openai.com/blog/musenet/ |archive-date=June 13, 2020 |access-date=June 8, 2020}} In pop culture, initial applications of this tool were used as early as 2020 for the internet psychological thriller Ben Drowned to create music for the titular character.{{Cite web |date=September 28, 2020 |title=Arcade Attack Podcast – September (4 of 4) 2020 - Alex Hall (Ben Drowned) - Interview |url=https://www.arcadeattack.co.uk/podcast-september-4-2020/ |url-status=live |archive-url=https://web.archive.org/web/20230203201108/https://www.arcadeattack.co.uk/podcast-september-4-2020/ |archive-date=February 3, 2023 |access-date=January 29, 2023 |website=Arcade Attack}}{{Cite web |author=Alexander Hall |date=June 25, 2020 |title=Tweets don't have titles and do not archive |work=X (formerly Twitter) |url=https://mobile.twitter.com/alexanderdhall/status/1276186528264392707 |url-status=live |archive-url=https://web.archive.org/web/20230203201107/https://mobile.twitter.com/alexanderdhall/status/1276186528264392707 |archive-date=February 3, 2023 |access-date=January 29, 2023}}

Released in 2020, Jukebox is an open-sourced algorithm to generate music with vocals. After training on 1.2 million samples, the system accepts a genre, artist, and a snippet of lyrics and outputs song samples. OpenAI stated the songs "show local musical coherence [and] follow traditional chord patterns" but acknowledged that the songs lack "familiar larger musical structures such as choruses that repeat" and that "there is a significant gap" between Jukebox and human-generated music. The Verge stated "It's technologically impressive, even if the results sound like mushy versions of songs that might feel familiar", while Business Insider stated "surprisingly, some of the resulting songs are catchy and sound legitimate".{{cite news |date=April 30, 2020 |title=OpenAI introduces Jukebox, a new AI model that generates genre-specific music |url=https://www.theverge.com/2020/4/30/21243038/openai-jukebox-model-raw-audio-lyrics-ai-generated-copyright |url-status=live |archive-url=https://web.archive.org/web/20200608043659/https://www.theverge.com/2020/4/30/21243038/openai-jukebox-model-raw-audio-lyrics-ai-generated-copyright |archive-date=June 8, 2020 |access-date=June 8, 2020 |work=The Verge}}{{cite news |last1=Stephen |first1=Bijan |date=April 30, 2020 |title=OpenAI introduces Jukebox, a new AI model that generates genre-specific music |url=https://www.businessinsider.com/jukebox-ai-music-generator-realistic-songs-machine-learning-algorithm-deepfakes-2020-5 |url-status=live |archive-url=https://web.archive.org/web/20200608043703/https://www.businessinsider.com/jukebox-ai-music-generator-realistic-songs-machine-learning-algorithm-deepfakes-2020-5 |archive-date=June 8, 2020 |access-date=June 8, 2020 |work=Business Insider |language=en}}{{cite web |date=April 30, 2020 |title=Jukebox |url=https://openai.com/blog/jukebox/ |url-status=live |archive-url=https://web.archive.org/web/20200608043703/https://openai.com/blog/jukebox/ |archive-date=June 8, 2020 |access-date=June 8, 2020 |publisher=OpenAI}}

In 2018, OpenAI launched the Debate Game, which teaches machines to debate toy problems in front of a human judge. The purpose is to research whether such an approach may assist in auditing AI decisions and in developing explainable AI.{{Cite news |last=Greene |first=Tristan |date=May 4, 2018 |title=OpenAI's Debate Game teaches you and your friends how to lie like robots |url=https://thenextweb.com/artificial-intelligence/2018/05/04/openais-debate-game-teaches-you-and-your-friends-how-to-lie-like-robots/ |url-status=live |archive-url=https://web.archive.org/web/20180505005129/https://thenextweb.com/artificial-intelligence/2018/05/04/openais-debate-game-teaches-you-and-your-friends-how-to-lie-like-robots/ |archive-date=May 5, 2018 |access-date=May 31, 2018 |work=The Next Web |language=en-US}}{{cite news |date=May 8, 2018 |title=Why Scientists Think AI Systems Should Debate Each Other |url=https://www.fastcompany.com/40569116/why-scientists-think-ai-systems-should-debate-each-other |url-status=live |archive-url=https://web.archive.org/web/20180519140829/https://www.fastcompany.com/40569116/why-scientists-think-ai-systems-should-debate-each-other |archive-date=May 19, 2018 |access-date=June 2, 2018 |work=Fast Company}}

Released in 2020, Microscope{{cite web |date=April 14, 2020 |title=OpenAI Microscope |url=https://openai.com/blog/microscope/ |url-status=live |archive-url=https://web.archive.org/web/20230203201911/https://openai.com/blog/microscope/ |archive-date=February 3, 2023 |access-date=March 27, 2021}} is a collection of visualizations of every significant layer and neuron of eight neural network models which are often studied in interpretability.{{Cite web |last=Johnson |first=Khari |date=April 14, 2020 |title=OpenAI launches Microscope to visualize the neurons in popular machine learning models |url=https://venturebeat.com/ai/openai-launches-microscope-to-visualize-the-neurons-in-popular-machine-learning-models/ |url-status=live |archive-url=https://web.archive.org/web/20230212154926/https://venturebeat.com/ai/openai-launches-microscope-to-visualize-the-neurons-in-popular-machine-learning-models/ |archive-date=February 12, 2023 |access-date=February 12, 2023 |website=VentureBeat |language=en-US}} Microscope was created to analyze the features that form inside these neural networks easily. The models included are AlexNet, VGG-19, different versions of Inception, and different versions of CLIP Resnet.{{Cite web |title=OpenAI Microscope |url=https://microscope.openai.com/models |url-status=live |archive-url=https://web.archive.org/web/20230203191623/https://microscope.openai.com/models |archive-date=February 3, 2023 |access-date=March 27, 2021 |website=OpenAI Microscope}}

{{main|ChatGPT}}File:OpenAI logo 2025 (symbol).svgLaunched in November 2022, ChatGPT is an artificial intelligence tool built on top of GPT-3 that provides a conversational interface that allows users to ask questions in natural language. The system then responds with an answer within seconds. ChatGPT reached 1 million users 5 days after its launch.{{cite web |date=December 5, 2022 |title=Mira Murati via Twitter |url=https://twitter.com/miramurati/status/1599796191243669504 |url-status=live |archive-url=https://web.archive.org/web/20221214090229/https://twitter.com/miramurati/status/1599796191243669504 |archive-date=December 14, 2022 |access-date=December 15, 2022 |publisher=Mira Murati}}{{Cite web |title=Pricing |url=https://openai.com/pricing |url-status=live |archive-url=https://web.archive.org/web/20230320175036/https://openai.com/pricing |archive-date=March 20, 2023 |access-date=March 20, 2023 |publisher=OpenAI}}

As of 2023, ChatGPT Plus is a GPT-4 backed version of ChatGPT{{Cite web |last=Edwards |first=Benj |date=March 14, 2023 |title=OpenAI's GPT-4 exhibits "human-level performance" on professional benchmarks |url=https://arstechnica.com/information-technology/2023/03/openai-announces-gpt-4-its-next-generation-ai-language-model/ |url-status=live |archive-url=https://web.archive.org/web/20230314225236/https://arstechnica.com/information-technology/2023/03/openai-announces-gpt-4-its-next-generation-ai-language-model/ |archive-date=March 14, 2023 |access-date=March 15, 2023 |website=Ars Technica}} available for a US$20 per month subscription fee{{Cite web |author=OpenAI |date=February 1, 2023 |title=Introducing ChatGPT Plus |url=https://openai.com/blog/chatgpt-plus |url-status=live |archive-url=https://web.archive.org/web/20230320005603/https://openai.com/blog/chatgpt-plus |archive-date=March 20, 2023 |access-date=March 20, 2023 |website=OpenAI Blog}} (the original version is backed by GPT-3.5).{{Cite web |author=OpenAI |title=OpenAI API |url=https://platform.openai.com/ |url-status=live |archive-url=https://web.archive.org/web/20230320023933/https://platform.openai.com/ |archive-date=March 20, 2023 |access-date=March 20, 2023 |website=platform.openai.com}} OpenAI also makes GPT-4 available to a select group of applicants through their GPT-4 API waitlist;{{Cite web |author=OpenAI |title=GPT-4 API waitlist |url=https://openai.com/waitlist/gpt-4-api |url-status=live |archive-url=https://web.archive.org/web/20230320174149/https://openai.com/waitlist/gpt-4-api |archive-date=March 20, 2023 |access-date=March 20, 2023 |website=openai.com}} after being accepted, an additional fee of US$0.03 per 1000 tokens in the initial text provided to the model ("prompt"), and US$0.06 per 1000 tokens that the model generates ("completion"), is charged for access to the version of the model with an 8192-token context window; for the 32768-token context window, the prices are doubled.{{Cite web |last=Wiggers |first=Kyle |date=February 1, 2023 |title=OpenAI launches ChatGPT Plus, starting at $20 per month |url=https://techcrunch.com/2023/02/01/openai-launches-chatgpt-plus-starting-at-20-per-month/ |access-date=February 12, 2023 |website=TechCrunch |language=en-US |archive-date=February 12, 2023 |archive-url=https://web.archive.org/web/20230212090719/https://techcrunch.com/2023/02/01/openai-launches-chatgpt-plus-starting-at-20-per-month/ |url-status=live }}

In May 2023, OpenAI launched a user interface for ChatGPT for the App Store on iOS and later in July 2023 for the Play Store on Android. The app supports chat history syncing and voice input (using Whisper, OpenAI's speech recognition model).{{Cite web |title=Introducing the ChatGPT app for iOS |url=https://openai.com/blog/introducing-the-chatgpt-app-for-ios |access-date=August 17, 2023 |website=openai.com |language=en-US |archive-date=June 5, 2023 |archive-url=https://web.archive.org/web/20230605064406/https://openai.com/blog/introducing-the-chatgpt-app-for-ios |url-status=live }}{{Cite web |last=Lawler |first=Richard |date=July 25, 2023 |title=ChatGPT for Android is now available |url=https://www.theverge.com/2023/7/25/23807012/chatgpt-android-ai-chatbot-openai-llm |access-date=August 17, 2023 |website=The Verge |language=en-US |archive-date=August 16, 2023 |archive-url=https://web.archive.org/web/20230816172328/https://www.theverge.com/2023/7/25/23807012/chatgpt-android-ai-chatbot-openai-llm |url-status=live }}{{Cite web |title=ChatGPT Android app FAQ {{!}} OpenAI Help Center |url=https://help.openai.com/en/articles/8142208-chatgpt-android-app-faq |access-date=August 17, 2023 |website=help.openai.com |language=en |archive-date=August 17, 2023 |archive-url=https://web.archive.org/web/20230817093628/https://help.openai.com/en/articles/8142208-chatgpt-android-app-faq |url-status=live }} In September 2023, OpenAI announced that ChatGPT "can now see, hear, and speak". ChatGPT Plus users can upload images, while mobile app users can talk to the chatbot.{{Cite web |title=ChatGPT can now see, hear, and speak |url=https://openai.com/blog/chatgpt-can-now-see-hear-and-speak |access-date=October 16, 2023 |website=openai.com |archive-date=November 7, 2023 |archive-url=https://web.archive.org/web/20231107080728/https://openai.com/blog/chatgpt-can-now-see-hear-and-speak |url-status=live }}{{Cite news |last=Roose |first=Kevin |date=September 27, 2023 |title=The New ChatGPT Can 'See' and 'Talk.' Here's What It's Like. |work=The New York Times |url=https://www.nytimes.com/2023/09/27/technology/new-chatgpt-can-see-hear.html |access-date=October 16, 2023 |via=NYTimes.com |archive-date=October 31, 2023 |archive-url=https://web.archive.org/web/20231031055345/https://www.nytimes.com/2023/09/27/technology/new-chatgpt-can-see-hear.html |url-status=live }}

In October 2023, OpenAI's latest image generation model, DALL-E 3, was integrated into ChatGPT Plus and ChatGPT Enterprise. The integration uses ChatGPT to write prompts for DALL-E guided by conversation with users.{{Cite web |last=David |first=Emilia |date=September 20, 2023 |title=OpenAI releases third version of DALL-E |url=https://www.theverge.com/2023/9/20/23881241/openai-dalle-third-version-generative-ai |access-date=September 23, 2023 |website=The Verge |language=en-US |archive-date=September 20, 2023 |archive-url=https://web.archive.org/web/20230920192429/https://www.theverge.com/2023/9/20/23881241/openai-dalle-third-version-generative-ai |url-status=live }}{{Cite news |last1=Metz |first1=Cade |last2=Hsu |first2=Tiffany |date=September 20, 2023 |title=ChatGPT Can Now Generate Images, Too |language=en-US |work=The New York Times |url=https://www.nytimes.com/2023/09/20/technology/chatgpt-dalle3-images-openai.html |access-date=September 23, 2023 |issn=0362-4331 |archive-date=September 23, 2023 |archive-url=https://web.archive.org/web/20230923154712/https://www.nytimes.com/2023/09/20/technology/chatgpt-dalle3-images-openai.html |url-status=live }}

OpenAI's GPT Store, initially slated for a 2023 launch, is now deferred to an undisclosed date in early 2024, attributed likely to the leadership changes in November following the initial announcement.{{cite news|last=Fried|first=Ina|date=December 1, 2023|title=Scoop: OpenAI delays launch of custom GPT store until early 2024|url=https://www.axios.com/2023/12/01/openai-delays-launch-custom-gpt-store-2024|work=Axios|access-date=December 3, 2023|archive-date=December 3, 2023|archive-url=https://web.archive.org/web/20231203003716/https://www.axios.com/2023/12/01/openai-delays-launch-custom-gpt-store-2024|url-status=live}}

Concerns about the energy consumption of generative AI, including ChatGPT, are rising. In September 2024, Microsoft entered a deal with Constellation Energy to reopen the Three Mile Island nuclear plant to supply power to its AI-driven data centers.{{Cite news |last=Halper |first=Evan |date=2024-09-20 |title=Microsoft deal would reopen Three Mile Island nuclear plant to power AI |url=https://www.washingtonpost.com/business/2024/09/20/microsoft-three-mile-island-nuclear-constellation/ |access-date=2024-10-07 |newspaper=Washington Post |language=en-US |issn=0190-8286}}

In December 2024, OpenAI launched a new feature allowing users to call ChatGPT for up to 15 minutes per month for free.{{Cite web |last=Samosa |first=Social |title=OpenAI launches free 15-minute phone calls with ChatGPT |url=https://www.socialsamosa.com/news-2/openai-launches-15-minute-phone-calls-chatgpt-8532973 |access-date=2024-12-20 |website=www.socialsamosa.com |language=en |archive-date=December 20, 2024 |archive-url=https://web.archive.org/web/20241220175224/https://www.socialsamosa.com/news-2/openai-launches-15-minute-phone-calls-chatgpt-8532973 |url-status=live }}{{Cite web |last=Field |first=Hayden |date=2024-12-18 |title=OpenAI makes ChatGPT available for phone calls and texts |url=https://www.cnbc.com/2024/12/18/openai-makes-chatgpt-available-for-phone-chats.html |access-date=2024-12-20 |website=CNBC |language=en |archive-date=December 20, 2024 |archive-url=https://web.archive.org/web/20241220135320/https://www.cnbc.com/2024/12/18/openai-makes-chatgpt-available-for-phone-chats.html |url-status=live }}

{{main|SearchGPT}}

SearchGPT, a prototype search engine developed by OpenAI, was unveiled on July 25, 2024, with an initial limited release to 10,000 test users. It combines traditional search engine features with generative AI capabilities.{{Cite web |last=Robison |first=Kylie |date=July 25, 2024 |title=OpenAI announces SearchGPT, its AI-powered search engine |url=https://www.theverge.com/2024/7/25/24205701/openai-searchgpt-ai-search-engine-google-perplexity-rival |access-date=July 27, 2024 |website=The Verge |language=en |archive-date=July 26, 2024 |archive-url=https://web.archive.org/web/20240726180050/https://www.theverge.com/2024/7/25/24205701/openai-searchgpt-ai-search-engine-google-perplexity-rival |url-status=live }}{{Cite web |last=Wiggers |first=Kyle |date=July 25, 2024 |title=With Google in its sights, OpenAI unveils SearchGPT |url=https://techcrunch.com/2024/07/25/with-google-in-its-sights-openai-unveils-searchgpt/ |access-date=July 26, 2024 |website=TechCrunch |language=en-US |archive-date=July 26, 2024 |archive-url=https://web.archive.org/web/20240726010010/https://techcrunch.com/2024/07/25/with-google-in-its-sights-openai-unveils-searchgpt/ |url-status=live }}

Stargate is a potential artificial intelligence supercomputer in development by Microsoft and OpenAI, in collaboration with Oracle, SoftBank, and MGX.{{Cite web |last1=Bajwa |last2=Simao |last3=Gregorio |first1=Arsheeya |first2=Paul |first3=David |date=March 29, 2024 |title=Microsoft, OpenAI plan $100 billion data-center project, media report says |website=Reuters |url=https://www.reuters.com/technology/microsoft-openai-planning-100-billion-data-center-project-information-reports-2024-03-29/ |access-date=June 6, 2024 |archive-date=June 20, 2024 |archive-url=https://web.archive.org/web/20240620203250/https://www.reuters.com/technology/microsoft-openai-planning-100-billion-data-center-project-information-reports-2024-03-29/ |url-status=live }}{{cite news |title=OpenAI, SoftBank, Oracle to invest $500 bln in AI, Trump says |url=https://www.reuters.com/technology/artificial-intelligence/openai-softbank-oracle-invest-500-bln-ai-trump-says-2025-01-21/ |publisher=Reuters}} Stargate is designed as part of a greater data center project, which could represent an investment of as much as $100 billion by Microsoft.{{Cite news |date=March 29, 2024 |title=Microsoft and OpenAI Plot $100 Billion Stargate AI Supercomputer |url=https://www.theinformation.com/articles/microsoft-and-openai-plot-100-billion-stargate-ai-supercomputer |access-date=June 6, 2024 |work=The Information}}

Stargate is reported to be part of a series of AI-related construction projects planned in the next few years by the companies Microsoft and OpenAI. The supercomputers will be constructed in five phases. The fourth phase should consist in a smaller OpenAI supercomputer, planned to launch around 2026. Stargate is the fifth and final phase of the program, and will take five and six years to complete and is slated to launch around 2028.

The artificial intelligence of Stargate is slated to be contained on millions of special server chips. The supercomputer's data center will be built in the US across 700 acres of land. It has a planned power consumption of 5 gigawatts, for which it could rely on nuclear energy. The name "Stargate" is a homage to the 1994 sci-fi film Stargate.

{{Further|Removal of Sam Altman from OpenAI}}

On November 17, 2023, Sam Altman was removed as CEO when its board of directors (composed of Helen Toner, Ilya Sutskever, Adam D'Angelo and Tasha McCauley) cited a lack of confidence in him. Chief Technology Officer Mira Murati took over as interim CEO. Greg Brockman, the president of OpenAI, was also removed as chairman of the board{{Cite web |title=OpenAI announces leadership transition |url=https://openai.com/blog/openai-announces-leadership-transition |access-date=November 17, 2023 |website=openai.com |language=en-US |archive-date=November 17, 2023 |archive-url=https://web.archive.org/web/20231117212221/https://openai.com/blog/openai-announces-leadership-transition |url-status=live }}{{cite news |last1=Montgomery |first1=Blake |last2=Anguiano |first2=Dani |date=November 17, 2023 |title=OpenAI fires co-founder and CEO Sam Altman for allegedly lying to company board |work=The Guardian |url=https://www.theguardian.com/technology/2023/nov/17/openai-ceo-sam-altman-fired |access-date=November 17, 2023 |archive-date=November 17, 2023 |archive-url=https://web.archive.org/web/20231117210649/https://www.theguardian.com/technology/2023/nov/17/openai-ceo-sam-altman-fired |url-status=live }} and resigned from the company's presidency shortly thereafter.{{cite web |last1=Peters |first1=Jay |title=OpenAI co-founder Greg Brockman is leaving, too |url=https://www.theverge.com/2023/11/17/23966277/openai-co-founder-greg-brockman-leaving |website=The Verge |language=en |date=November 18, 2023 |access-date=November 18, 2023 |archive-date=November 18, 2023 |archive-url=https://web.archive.org/web/20231118003259/https://www.theverge.com/2023/11/17/23966277/openai-co-founder-greg-brockman-leaving |url-status=live }} Three senior OpenAI researchers subsequently resigned: director of research and GPT-4 lead Jakub Pachocki, head of AI risk {{ill|Aleksander Mądry|pl}}, and researcher Szymon Sidor.{{cite news |title=Three Senior OpenAI Researchers Resign as Crisis Deepens |work=The Information |url=https://www.theinformation.com/articles/three-senior-openai-researchers-resign-as-crisis-deepens |access-date=November 18, 2023 |archive-date=November 18, 2023 |archive-url=https://web.archive.org/web/20231118065056/https://www.theinformation.com/articles/three-senior-openai-researchers-resign-as-crisis-deepens |url-status=live }}{{cite news |last1=Edwards |first1=Benj |title=Details emerge of surprise board coup that ousted CEO Sam Altman at OpenAI |url=https://arstechnica.com/information-technology/2023/11/report-sutskever-led-board-coup-at-openai-that-ousted-altman-over-ai-safety-concerns/ |access-date=November 19, 2023 |work=Ars Technica |date=November 18, 2023 |language=en-us |archive-date=November 19, 2023 |archive-url=https://web.archive.org/web/20231119020114/https://arstechnica.com/information-technology/2023/11/report-sutskever-led-board-coup-at-openai-that-ousted-altman-over-ai-safety-concerns/ |url-status=live }}

On November 18, 2023, there were reportedly talks of Altman returning as CEO amid pressure placed upon the board by investors such as Microsoft and Thrive Capital, who objected to Altman's departure.{{cite news |author1=Keach Hagey |author2=Berber Jin |author3=Deepa Seetharaman |title=OpenAI Investors Trying to Get Sam Altman Back as CEO After Sudden Firing |url=https://www.wsj.com/tech/openai-trying-to-get-sam-altman-back-4b728049 |website=WSJ |access-date=November 19, 2023 |archive-date=November 18, 2023 |archive-url=https://web.archive.org/web/20231118233153/https://www.wsj.com/tech/openai-trying-to-get-sam-altman-back-4b728049 |url-status=live }} Although Altman himself spoke in favor of returning to OpenAI, he has since stated that he considered starting a new company and bringing former OpenAI employees with him if talks to reinstate him didn't work out.{{cite web |last1=Metz |first1=Cade |last2=Isaac |first2=Mike |last3=Griffith |first3=Erin |title=Sam Altman Is Said to Be Discussing Return to OpenAI With Company's Board |url=https://www.nytimes.com/2023/11/18/technology/sam-altman-openai-board.html?smid=nytcore-ios-share&referringSource=articleShare |website=The New York Times |date=November 19, 2023 |access-date=November 19, 2023 |archive-date=November 19, 2023 |archive-url=https://web.archive.org/web/20231119014720/https://www.nytimes.com/2023/11/18/technology/sam-altman-openai-board.html?smid=nytcore-ios-share&referringSource=articleShare |url-status=live }} The board members agreed "in principle" to resign if Altman returned.{{cite web |last1=Patel |first1=Nilay |title=OpenAI board in discussions with Sam Altman to return as CEO |url=https://www.theverge.com/2023/11/18/23967199/breaking-openai-board-in-discussions-with-sam-altman-to-return-as-ceo |website=The Verge |language=en |date=November 18, 2023 |access-date=November 19, 2023 |archive-date=November 18, 2023 |archive-url=https://web.archive.org/web/20231118225333/https://www.theverge.com/2023/11/18/23967199/breaking-openai-board-in-discussions-with-sam-altman-to-return-as-ceo |url-status=live }} On November 19, 2023, negotiations with Altman to return failed and Murati was replaced by Emmett Shear as interim CEO.{{cite web |last1=Heath |first1=Alex |title=The deal to bring Sam Altman back to OpenAI has fallen apart |url=https://www.theverge.com/2023/11/20/23967515/sam-altman-openai-board-fired-new-ceo |website=The Verge |language=en |date=November 19, 2023 |access-date=November 20, 2023 |archive-date=November 20, 2023 |archive-url=https://web.archive.org/web/20231120053238/https://www.theverge.com/2023/11/20/23967515/sam-altman-openai-board-fired-new-ceo |url-status=live }} The board initially contacted Anthropic CEO Dario Amodei (a former OpenAI executive) about replacing Altman, and proposed a merger of the two companies, but both offers were declined.{{cite web |last1=Dastin |first1=Jeffrey |title=OpenAI's board approached Anthropic CEO about top job and merger |url=https://www.reuters.com/technology/openais-board-approached-anthropic-ceo-about-top-job-merger-sources-2023-11-21/ |website=Reuters |language=en |date=November 21, 2023 |access-date=November 21, 2023 |archive-date=November 21, 2023 |archive-url=https://web.archive.org/web/20231121203624/https://www.reuters.com/technology/openais-board-approached-anthropic-ceo-about-top-job-merger-sources-2023-11-21/ |url-status=live }}

On November 20, 2023, Microsoft CEO Satya Nadella announced Altman and Brockman would be joining Microsoft to lead a new advanced AI research team, but added that they were still committed to OpenAI despite recent events.{{cite web |last1=Warren |first1=Tom |title=Microsoft hires former OpenAI CEO Sam Altman |url=https://www.theverge.com/2023/11/20/23968829/microsoft-hires-sam-altman-greg-brockman-employees-openai |website=The Verge |language=en |date=November 20, 2023 |access-date=November 20, 2023 |archive-date=November 20, 2023 |archive-url=https://web.archive.org/web/20231120133123/https://www.theverge.com/2023/11/20/23968829/microsoft-hires-sam-altman-greg-brockman-employees-openai |url-status=live }} Before the partnership with Microsoft was finalized, Altman gave the board another opportunity to negotiate with him.{{cite web |last1=Patel |first1=Nilay |title=Sam Altman is still trying to return as OpenAI CEO |url=https://www.theverge.com/2023/11/20/23969586/sam-altman-plotting-return-open-ai-microsoft |website=The Verge |language=en |date=November 20, 2023 |access-date=November 20, 2023 |archive-date=November 20, 2023 |archive-url=https://web.archive.org/web/20231120191138/https://www.theverge.com/2023/11/20/23969586/sam-altman-plotting-return-open-ai-microsoft |url-status=live }} About 738 of OpenAI's 770 employees, including Murati and Sutskever, signed an open letter stating they would quit their jobs and join Microsoft if the board did not rehire Altman and then resign.{{cite web |title=OpenAI Staff Near Total Mutiny With Threat to Jump to Microsoft |url=https://news.bloomberglaw.com/us-law-week/openai-staff-threaten-to-go-to-microsoft-if-board-doesnt-quit |website=Bloomberg |date=November 20, 2023 |language=en |access-date=November 20, 2023 |archive-date=November 20, 2023 |archive-url=https://web.archive.org/web/20231120201119/https://news.bloomberglaw.com/us-law-week/openai-staff-threaten-to-go-to-microsoft-if-board-doesnt-quit |url-status=live }}{{cite magazine |last1=Knight |first1=Will |title=OpenAI Staff Threaten to Quit Unless Board Resigns |url=https://www.wired.com/story/openai-staff-walk-protest-sam-altman/ |magazine=Wired |access-date=November 20, 2023 |archive-date=November 20, 2023 |archive-url=https://web.archive.org/web/20231120150129/https://www.wired.com/story/openai-staff-walk-protest-sam-altman/ |url-status=live }} This prompted OpenAI investors to consider legal action against the board as well.{{cite web |last1=Tong |first1=Anna |last2=Hu |first2=Krystal |last3=Tong |first3=Anna |last4=Hu |first4=Krystal |title=Exclusive: OpenAI investors considering suing the board after CEO's abrupt firing |url=https://www.reuters.com/technology/openai-investors-considering-suing-board-after-ceos-abrupt-firing-sources-2023-11-20/ |website=Reuters |language=en |date=November 20, 2023 |access-date=November 20, 2023 |archive-date=November 20, 2023 |archive-url=https://web.archive.org/web/20231120230652/https://www.reuters.com/technology/openai-investors-considering-suing-board-after-ceos-abrupt-firing-sources-2023-11-20/ |url-status=live }} In response, OpenAI management sent an internal memo to employees stating that negotiations with Altman and the board had resumed and would take some time.{{cite web |last1=Lawler |first1=Richard |title=OpenAI exec to employees: "our number one goal remains to reunify OpenAI." |url=https://www.theverge.com/2023/11/21/23970550/openai-exec-to-employees-our-number-one-goal-remains-to-reunify-openai |website=The Verge |language=en |date=November 21, 2023 |access-date=November 21, 2023 |archive-date=November 21, 2023 |archive-url=https://web.archive.org/web/20231121180928/https://www.theverge.com/2023/11/21/23970550/openai-exec-to-employees-our-number-one-goal-remains-to-reunify-openai |url-status=live }}

{{Anchor|Q*}}

On November 21, 2023, after continued negotiations, Altman and Brockman returned to the company in their prior roles along with a reconstructed board made up of new members Bret Taylor (as chairman) and Lawrence Summers, with D'Angelo remaining.{{cite web |last1=Heath |first1=Alex |title=Breaking: Sam Altman to return as CEO of OpenAI |url=https://www.theverge.com/2023/11/22/23967223/sam-altman-returns-ceo-open-ai |website=The Verge |language=en |date=November 22, 2023 |access-date=November 22, 2023 |archive-date=November 22, 2023 |archive-url=https://web.archive.org/web/20231122060733/https://www.theverge.com/2023/11/22/23967223/sam-altman-returns-ceo-open-ai |url-status=live }} On November 22, 2023, emerging reports suggested that Sam Altman's dismissal from OpenAI may have been linked to his alleged mishandling of a significant breakthrough in the organization's secretive project codenamed Q*. According to sources within OpenAI, Q* is aimed at developing AI capabilities in logical and mathematical reasoning, and reportedly involves performing math on the level of grade-school students.{{cite news|title=Exclusive: OpenAI researchers warned board of AI breakthrough ahead of CEO ouster, sources say |author=Anna Tong|author2=Jeffrey Dastin|author3=Krystal Hu|work=Reuters|url=https://www.reuters.com/technology/sam-altmans-ouster-openai-was-precipitated-by-letter-board-about-ai-breakthrough-2023-11-22|date=November 22, 2023|quote=Some at OpenAI believe Q* (pronounced Q-Star) could be a breakthrough in the startup's search for what's known as artificial general intelligence (AGI), one of the people told Reuters. OpenAI defines AGI as autonomous systems that surpass humans in most economically valuable tasks.|access-date=November 23, 2023|archive-date=November 23, 2023 |archive-url=https://web.archive.org/web/20231123045247/https://www.reuters.com/technology/sam-altmans-ouster-openai-was-precipitated-by-letter-board-about-ai-breakthrough-2023-11-22/|url-status=live}}{{Cite web |title=openai/prm800k |date=January 16, 2024 |via=GitHub |url=https://github.com/openai/prm800k|access-date=December 4, 2023|archive-date=November 24, 2023|archive-url=https://web.archive.org/web/20231124210953/https://github.com/openai/prm800k|url-status=live}}{{Cite arXiv |title=Let's Verify Step by Step|eprint=2305.20050 |last1=Lightman |first1=Hunter |last2=Kosaraju |first2=Vineet |last3=Burda |first3=Yura |last4=Edwards |first4=Harri |last5=Baker |first5=Bowen |last6=Lee |first6=Teddy |last7=Leike |first7=Jan |last8=Schulman |first8=John |last9=Sutskever |first9=Ilya |last10=Cobbe |first10=Karl |date=2023 |class=cs.LG }} Concerns about Altman's response to this development, specifically regarding the discovery's potential safety implications, were reportedly raised with the company's board shortly before Altman's firing.{{cite web | last1=Tong | first1=Anna | last2=Dastin | first2=Jeffrey | last3=Hu | first3=Krystal | title=Exclusive: OpenAI researchers warned board of AI breakthrough ahead of CEO ouster, sources say | website=Reuters | date=November 23, 2023 | url=https://www.reuters.com/technology/sam-altmans-ouster-openai-was-precipitated-by-letter-board-about-ai-breakthrough-2023-11-22/ | access-date=November 23, 2023 | archive-date=December 11, 2023 | archive-url=https://web.archive.org/web/20231211175826/https://www.reuters.com/technology/sam-altmans-ouster-openai-was-precipitated-by-letter-board-about-ai-breakthrough-2023-11-22/ | url-status=live }} On November 29, 2023, OpenAI announced that an anonymous Microsoft employee had joined the board as a non-voting member to observe the company's operations;{{cite web |last1=Heath |first1=Alex |title=Microsoft joins OpenAI's board with Sam Altman officially back as CEO |url=https://www.theverge.com/2023/11/29/23981848/sam-altman-back-open-ai-ceo-microsoft-board |website=The Verge |language=en |date=November 30, 2023 |access-date=December 14, 2023 |archive-date=December 14, 2023 |archive-url=https://web.archive.org/web/20231214033303/https://www.theverge.com/2023/11/29/23981848/sam-altman-back-open-ai-ceo-microsoft-board |url-status=live }} Microsoft resigned from the board in July 2024.{{Cite web |title=Microsoft ditches OpenAI board observer seat amid regulatory scrutiny |url=https://www.channelnewsasia.com/business/microsoft-ditches-openai-board-observer-seat-amid-regulatory-scrutiny-4469086 |access-date=July 10, 2024 |website=CNA |language=en |archive-date=July 11, 2024 |archive-url=https://web.archive.org/web/20240711100155/https://www.channelnewsasia.com/business/microsoft-ditches-openai-board-observer-seat-amid-regulatory-scrutiny-4469086 |url-status=live }}

{{Broader|Content moderation and working conditions}}

In January 2023, OpenAI has been criticized for outsourcing the annotation of data sets to Sama, a company based in San Francisco that employed workers in Kenya. These annotations were used to train an AI model to detect toxicity, which could then be used to moderate toxic content, notably from ChatGPT's training data and outputs. However, these pieces of text usually contained detailed descriptions of various types of violence, including sexual violence. The investigation uncovered that OpenAI began sending snippets of data to Sama as early as November 2021. The four Sama employees interviewed by Time described themselves as mentally scarred. OpenAI paid Sama $12.50 per hour of work, and Sama was redistributing the equivalent of between $1.32 and $2.00 per hour post-tax to its annotators. Sama's spokesperson said that the $12.50 was also covering other implicit costs, among which were infrastructure expenses, quality assurance and management.{{cite magazine |last1=Perrigo |first1=Billy |title=Exclusive: The $2 Per Hour Workers Who Made ChatGPT Safer |url=https://time.com/6247678/openai-chatgpt-kenya-workers/ |access-date=August 5, 2023 |magazine=Time |date=January 18, 2023 |language=en |archive-date=January 19, 2023 |archive-url=https://web.archive.org/web/20230119152814/https://time.com/6247678/openai-chatgpt-kenya-workers/ |url-status=live }}

In March 2023, the company was also criticized for disclosing particularly few technical details about products like GPT-4, contradicting its initial commitment to openness and making it harder for independent researchers to replicate its work and develop safeguards. OpenAI cited competitiveness and safety concerns to justify this strategic turn. OpenAI's former chief scientist Ilya Sutskever argued in 2023 that open-sourcing increasingly capable models was increasingly risky, and that the safety reasons for not open-sourcing the most potent AI models would become "obvious" in a few years.{{cite news |last1=Vincent |first1=James |title=OpenAI co-founder on company's past approach to openly sharing research: "We were wrong" |url=https://www.theverge.com/2023/3/15/23640180/openai-gpt-4-launch-closed-research-ilya-sutskever-interview |access-date=August 20, 2023 |work=The Verge |date=March 15, 2023 |archive-date=March 17, 2023 |archive-url=https://web.archive.org/web/20230317210900/https://www.theverge.com/2023/3/15/23640180/openai-gpt-4-launch-closed-research-ilya-sutskever-interview |url-status=live }}

On May 17, 2024, a Vox article reported that OpenAI was asking departing employees to sign a lifelong non-disparagement agreement forbidding them from criticizing OpenAI or acknowledging the existence of the agreement. Daniel Kokotajlo, a former employee, publicly stated that he forfeited his vested equity in OpenAI in order to leave without signing the agreement.{{Cite web |last=Piper |first=Kelsey |date=May 17, 2024 |title=ChatGPT can talk, but OpenAI employees sure can't |url=https://www.vox.com/future-perfect/2024/5/17/24158478/openai-departures-sam-altman-employees-chatgpt-release |access-date=May 18, 2024 |website=Vox |language=en |archive-date=May 18, 2024 |archive-url=https://web.archive.org/web/20240518221431/https://www.vox.com/future-perfect/2024/5/17/24158478/openai-departures-sam-altman-employees-chatgpt-release |url-status=live }}{{Cite web |last=Christian |first=Jon |date=May 18, 2024 |title=OpenAI Employees Forced to Sign NDA Preventing Them From Ever Criticizing Company |url=https://futurism.com/the-byte/openai-nda-criticism |access-date=May 18, 2024 |website=Futurism |archive-date=May 18, 2024 |archive-url=https://web.archive.org/web/20240518221433/https://futurism.com/the-byte/openai-nda-criticism |url-status=live }} Sam Altman stated that he was unaware of the equity cancellation provision, and that OpenAI never enforced it to cancel any employee's vested equity.{{Cite web |last=Getahun |first=Hannah |title=Sam Altman addresses 'potential equity cancellation' in OpenAI exit agreements after 2 high-profile departures |url=https://www.businessinsider.com/sam-altman-openai-nda-clause-vested-equity-ilya-sutskever-2024-5 |access-date=May 19, 2024 |website=Business Insider |language=en-US |archive-date=May 19, 2024 |archive-url=https://web.archive.org/web/20240519130546/https://www.businessinsider.com/sam-altman-openai-nda-clause-vested-equity-ilya-sutskever-2024-5 |url-status=live }} Vox published leaked documents and emails challenging this claim.{{Cite web |last=Piper |first=Kelsey |date=May 22, 2024 |title=Leaked OpenAI documents reveal aggressive tactics toward former employees |url=https://www.vox.com/future-perfect/351132/openai-vested-equity-nda-sam-altman-documents-employees |access-date=June 2, 2024 |website=Vox |language=en-US |archive-date=June 1, 2024 |archive-url=https://web.archive.org/web/20240601161425/https://www.vox.com/future-perfect/351132/openai-vested-equity-nda-sam-altman-documents-employees |url-status=live }} On May 23, 2024, OpenAI sent a memo releasing former employees from the agreement.{{Cite web |last=Field |first=Hayden |date=May 24, 2024 |title=OpenAI sends internal memo releasing former employees from controversial exit agreements |url=https://www.cnbc.com/2024/05/24/openai-sends-internal-memo-releasing-former-employees-from-non-disparagement-agreements-sam-altman.html |access-date=June 2, 2024 |website=CNBC |language=en |archive-date=May 30, 2024 |archive-url=https://web.archive.org/web/20240530032655/https://www.cnbc.com/2024/05/24/openai-sends-internal-memo-releasing-former-employees-from-non-disparagement-agreements-sam-altman.html |url-status=live }}

OpenAI, Inc. was originally designed as a nonprofit in order to ensure that AGI "benefits all of humanity" rather than "the private gain of any person". In 2019, it created OpenAI Global, LLC, a capped-profit subsidiary controlled by the nonprofit. In December 2024, OpenAI proposed a restructuring plan to convert the capped-profit into a Delaware-based public benefit corporation (PBC), and to release it from the control of the nonprofit. The nonprofit would sell its control and other assets, getting equity in return, and would use it to fund and pursue separate charitable projects, including in science and education. OpenAI's leadership described the change as necessary to secure additional investments, and claimed that the nonprofit's founding mission to ensure AGI "benefits all of humanity" would be better fulfilled.{{Cite web |last=Booth |first=Harry |date=2025-04-24 |title=OpenAI Wants to Go For-Profit. Experts Say Regulators Should Step In |url=https://time.com/7279977/openai-for-profit-letter-elon-musk/ |access-date=2025-04-25 |website=TIME |language=en}}

The plan has been criticized by experts and former employees. A legal letter named "[https://notforprivategain.org/ Not For Private Gain]" asked the attorneys general of California and Delaware to intervene, stating that the restructuring is illegal and would remove governance safeguards from the nonprofit and the attorneys general.{{Cite web |last=Goldman |first=Sharon |title=Ex-OpenAI employees sign open letter to California AG: For-profit pivot poses ‘palpable threat’ to nonprofit mission |url=https://fortune.com/article/ex-openai-employees-california-ag-for-profit-pivot-threat-nonprofit-mission/ |access-date=2025-04-25 |website=Fortune |language=en}} The letter argues that OpenAI's complex structure was deliberately designed to remain accountable to its mission, without the conflicting pressure of maximizing profits. It contends that the nonprofit is best positioned to advance its mission of ensuring AGI benefits all of humanity by continuing to control OpenAI Global, LLC, whatever the amount of equity that it could get in exchange.{{Cite web |last=Piper |first=Kelsey |date=2025-04-24 |title=OpenAI’s nonprofit structure was supposed to protect you. What went wrong? |url=https://www.vox.com/future-perfect/410261/openai-non-profit-transition-letter-sam-altman-artificial-intelligence |access-date=2025-04-26 |website=Vox |language=en-US}} PBCs can choose how they balance their mission with profit-making. Controlling shareholders have a large influence on how closely a PBC sticks to its mission.{{Cite news |date=January 2, 2025 |title=OpenAI outlines new for-profit structure in bid to stay ahead in costly AI race |url=https://www.reuters.com/technology/artificial-intelligence/openai-lays-out-plan-shift-new-for-profit-structure-2024-12-27/ |work=Reuters}}

Legally, under nonprofit law, assets dedicated to a charitable purpose must continue to serve that purpose. To change its purpose, OpenAI would have to prove that its current purposes have become unlawful, impossible, impracticable, or wasteful.{{Cite news |date=February 13, 2025 |title=OpenAI Conversion Sheds Nonprofit Purpose Without Justification |url=https://news.bloomberglaw.com/us-law-week/openai-conversion-sheds-nonprofit-purpose-without-justification}} Elon Musk, who had initiated a lawsuit against OpenAI and Altman in August 2024 alleging the company violated contract provisions by prioritizing profit over its mission, reportedly leveraged this lawsuit to stop the restructuring plan. On February 10, 2025, a consortium of investors led by Elon Musk submitted a $97.4 billion unsolicited bid to buy the nonprofit that controls OpenAI, declaring willingness to match or exceed any better offer.{{cite news |last1=Hammond |first1=George |date=10 February 2025 |title=Elon Musk-led consortium offers $100bn to take control of OpenAI |url=https://www.ft.com/content/3a673ed2-26d5-47af-9028-8af7d742c2e7 |access-date=11 February 2025 |work=Financial Times}}{{Cite news |last1=Toonkel |first1=Jessica |last2=Jin |first2=Berber |date=February 10, 2025 |title=Elon Musk-Led Group Makes $97.4 Billion Bid for Control of OpenAI |url=https://www.wsj.com/tech/elon-musk-openai-bid-4af12827 |newspaper=The Wall Street Journal}} The offer was rejected on 14 February 2025, with OpenAI stating that it was not for sale,{{cite news |date=14 February 2025 |title=OpenAI rejects $97.4bn Musk bid and says company is not for sale |url=https://www.theguardian.com/technology/2025/feb/14/openai-elon-musk |access-date=15 February 2025 |website=The Guardian}} but the offer complicated Altman's restructuring plan by suggesting a lower bar for how much the nonprofit should be valued.

In May 2025, the nonprofit's board chairman Bret Taylor announced that the nonprofit would renounce plans to cede control after outside pressure. The capped-profit still plans to transition to a PBC,{{Cite web |last=Field |first=Hayden |date=2025-05-05 |title=OpenAI says nonprofit will retain control of company, bowing to outside pressure |url=https://www.cnbc.com/2025/05/05/openai-says-nonprofit-retain-control-of-company-bowing-to-pressure.html |access-date=2025-05-05 |website=CNBC |language=en}} which critics said would diminish the nonprofit's control.{{Cite news |date=2025-05-15 |title=Group that opposed OpenAI's restructuring raises concerns about new revamp plan |url=https://www.reuters.com/business/group-that-opposed-openais-restructuring-raises-concerns-about-new-revamp-plan-2025-05-15/ |work=Reuters}}

OpenAI was sued for copyright infringement by authors Sarah Silverman, Matthew Butterick, Paul Tremblay and Mona Awad in July 2023.{{Cite web |last=Belanger |first=Ashley |date=July 10, 2023 |title=Sarah Silverman sues OpenAI, Meta for being "industrial-strength plagiarists" |url=https://arstechnica.com/information-technology/2023/07/book-authors-sue-openai-and-meta-over-text-used-to-train-ai/ |access-date=September 21, 2023 |website=Ars Technica |language=en-us |archive-date=September 21, 2023 |archive-url=https://web.archive.org/web/20230921134950/https://arstechnica.com/information-technology/2023/07/book-authors-sue-openai-and-meta-over-text-used-to-train-ai/ |url-status=live }}{{Cite web |last=Krithika |first=K. L. |date=August 21, 2023 |title=Legal Challenges Surround OpenAI: A Closer Look at the Lawsuits |url=https://analyticsindiamag.com/all-the-lawsuits-filed-against-openai/ |url-status=live |archive-url=https://web.archive.org/web/20230823225353/https://analyticsindiamag.com/all-the-lawsuits-filed-against-openai/ |archive-date=August 23, 2023 |access-date=August 23, 2023 |website=Analytics India Magazine |language=en-US}}{{Cite web |last=Abshire |first=Elisha |date=July 6, 2023 |title=OpenAI faces copyright lawsuit from authors Mona Awad and Paul Tremblay |url=https://dailyai.com/2023/07/openai-faces-copyright-lawsuit-from-authors-mona-awad-and-paul-tremblay/ |access-date=July 19, 2023 |website=Dailyai.com |language=en-uk |archive-date=July 18, 2023 |archive-url=https://web.archive.org/web/20230718091105/https://dailyai.com/2023/07/openai-faces-copyright-lawsuit-from-authors-mona-awad-and-paul-tremblay/ |url-status=live }} In September 2023, 17 authors, including George R. R. Martin, John Grisham, Jodi Picoult and Jonathan Franzen, joined the Authors Guild in filing a class action lawsuit against OpenAI, alleging that the company's technology was illegally using their copyrighted work.{{Cite web |last=Belanger |first=Ashley |date=September 20, 2023 |title=Grisham, Martin join authors suing OpenAI: "There is nothing fair about this" [Updated] |url=https://arstechnica.com/tech-policy/2023/09/george-r-r-martin-joins-authors-suing-openai-over-copyright-infringement/ |access-date=September 21, 2023 |website=Ars Technica |language=en-us |archive-date=September 21, 2023 |archive-url=https://web.archive.org/web/20230921164338/https://arstechnica.com/tech-policy/2023/09/george-r-r-martin-joins-authors-suing-openai-over-copyright-infringement/ |url-status=live }}{{Cite web |last=Korn |first=Jennifer |date=September 20, 2023 |title=George R. R. Martin, Jodi Picoult and other famous writers join Authors Guild in class action lawsuit against OpenAI |url=https://www.cnn.com/2023/09/20/tech/authors-guild-openai-lawsuit/index.html |url-status=live |archive-url=https://web.archive.org/web/20230921141901/https://www.cnn.com/2023/09/20/tech/authors-guild-openai-lawsuit/index.html |archive-date=September 21, 2023 |access-date=September 21, 2023 |website=CNN Business |language=en}} The New York Times also sued the company in late December 2023.{{cite news |date=December 27, 2023 |title=NY Times sues OpenAI, Microsoft for infringing copyrighted works |url=https://www.reuters.com/legal/transactional/ny-times-sues-openai-microsoft-infringing-copyrighted-work-2023-12-27 |access-date=December 27, 2023 |url-status=live |archive-url=https://web.archive.org/web/20231230163134/https://www.reuters.com/legal/transactional/ny-times-sues-openai-microsoft-infringing-copyrighted-work-2023-12-27/ |archive-date=December 30, 2023 |work=Reuters}} In May 2024 it was revealed that OpenAI had destroyed its Books1 and Books2 training datasets, which were used in the training of GPT-3, and which the Authors Guild believed to have contained over 100,000 copyrighted books.{{Cite web |author1=Darius Rafieyan |author2=Hasan Chowdhury |title=OpenAI destroyed a trove of books used to train AI models. The employees who collected the data are gone. |url=https://www.businessinsider.com/openai-destroyed-ai-training-datasets-lawsuit-authors-books-copyright-2024-5 |access-date=May 8, 2024 |website=Business Insider |language=en-US |quote=…Authors Guild said … the datasets likely contained "more than 100,000 published books" … central to its allegations that OpenAI used copyrighted materials to train AI models |archive-date=May 7, 2024 |archive-url=https://web.archive.org/web/20240507235039/https://www.businessinsider.com/openai-destroyed-ai-training-datasets-lawsuit-authors-books-copyright-2024-5 |url-status=live }}

In 2021, OpenAI developed a speech recognition tool called Whisper. OpenAI used it to transcribe more than one million hours of YouTube videos into text for training GPT-4. The automated transcription of YouTube videos raised concerns within OpenAI employees regarding potential violations of YouTube's terms of service, which prohibit the use of videos for applications independent of the platform, as well as any type of automated access to its videos. Despite these concerns, the project proceeded with notable involvement from OpenAI's president, Greg Brockman. The resulting dataset proved instrumental in training GPT-4.{{Cite news |last1=Metz |first1=Cade |last2=Kang |first2=Cecilia |last3=Frenkel |first3=Sheera |last4=Thompson |first4=Stuart A. |last5=Grant |first5=Nico |date=April 6, 2024 |title=How Tech Giants Cut Corners to Harvest Data for A.I. |url=https://www.nytimes.com/2024/04/06/technology/tech-giants-harvest-data-artificial-intelligence.html |work=The New York Times |access-date=May 21, 2024 |archive-date=May 5, 2024 |archive-url=https://web.archive.org/web/20240505040123/https://www.nytimes.com/2024/04/06/technology/tech-giants-harvest-data-artificial-intelligence.html |url-status=live }}

In February 2024, The Intercept as well as Raw Story and Alternate Media Inc. filed lawsuit against OpenAI on copyright litigation ground.{{Cite news |last=Brittain |first=Blake |date=February 29, 2024 |title=OpenAI hit with new lawsuits from news outlets over AI training |url=https://www.reuters.com/legal/litigation/openai-hit-with-new-lawsuits-news-outlets-over-ai-training-2024-02-28/ |access-date=March 24, 2024 |work=Reuters}}[https://fingfx.thomsonreuters.com/gfx/legaldocs/xmpjrjwjrpr/OPENAI%20RAW%20STORY%20LAWSUIT%20intercept.pdf OpenAI RAW STORY LAWSUIT INTERCEPT] {{Webarchive|url=https://web.archive.org/web/20240328115412/https://fingfx.thomsonreuters.com/gfx/legaldocs/xmpjrjwjrpr/OPENAI%20RAW%20STORY%20LAWSUIT%20intercept.pdf|date=March 28, 2024}} - from Reuters The lawsuit is said to have charted a new legal strategy for digital-only publishers to sue OpenAI.{{Cite web |title=The Intercept charts a new legal strategy for digital publishers suing OpenAI |url=https://www.niemanlab.org/2024/03/the-intercept-charts-a-new-legal-strategy-for-digital-publishers-suing-openai/ |url-status=live |archive-url=https://web.archive.org/web/20240328115412/https://www.niemanlab.org/2024/03/the-intercept-charts-a-new-legal-strategy-for-digital-publishers-suing-openai/ |archive-date=March 28, 2024 |access-date=March 28, 2024 |website=Nieman Lab}}

On April 30, 2024, eight newspapers filed a lawsuit in the Southern District of New York against OpenAI and Microsoft, claiming illegal harvesting of their copyrighted articles. The suing publications included The Mercury News, The Denver Post, The Orange County Register, St. Paul Pioneer Press, Chicago Tribune, Orlando Sentinel, Sun Sentinel, and New York Daily News.{{cite news |last1=Baron |first1=Ethan |date=April 30, 2024 |title=Mercury News and other papers sue Microsoft, OpenAI over the new artificial intelligence |url=https://www.eastbaytimes.com/2024/04/30/mercury-news-and-other-papers-sue-microsoft-openai-over-the-new-artificial-intelligence/ |access-date=April 30, 2024 |work=East Bay Times |archive-date=April 30, 2024 |archive-url=https://web.archive.org/web/20240430162841/https://www.eastbaytimes.com/2024/04/30/mercury-news-and-other-papers-sue-microsoft-openai-over-the-new-artificial-intelligence/ |url-status=live }}

In April 2023, the EU's European Data Protection Board (EDPB) formed a dedicated task force on ChatGPT "to foster cooperation and to exchange information on possible enforcement actions conducted by data protection authorities" based on the "enforcement action undertaken by the Italian data protection authority against Open AI about the Chat GPT service".{{cite web |title=EDPB resolves dispute on transfers by Meta and creates task force on Chat GPT |url=https://edpb.europa.eu/news/news/2023/edpb-resolves-dispute-transfers-meta-and-creates-task-force-chat-gpt_en |website=EDPB resolves dispute on transfers by Meta and creates task force on Chat GPT |access-date=November 22, 2023 |archive-date=November 22, 2023 |archive-url=https://web.archive.org/web/20231122215754/https://edpb.europa.eu/news/news/2023/edpb-resolves-dispute-transfers-meta-and-creates-task-force-chat-gpt_en |url-status=live }}

In late April 2024 NOYB filed a complaint with the Austrian Datenschutzbehörde against OpenAI for violating the European General Data Protection Regulation. A text created with ChatGPT gave a false date of birth for a living person without giving the individual the option to see the personal data used in the process. A request to correct the mistake was denied. Additionally, neither the recipients of ChatGPT's work nor the sources used, could be made available, OpenAI claimed.{{Cite web |title=ChatGPT verbreitet falsche Infos über Personen – und OpenAI kann nichts tun |url=https://noyb.eu/de/chatgpt-provides-false-information-about-people-and-openai-cant-correct-it |access-date=April 29, 2024 |website=noyb.eu |archive-date=April 29, 2024 |archive-url=https://web.archive.org/web/20240429054826/https://noyb.eu/de/chatgpt-provides-false-information-about-people-and-openai-cant-correct-it |url-status=live }}

OpenAI was criticized for lifting its ban on using ChatGPT for "military and warfare". Up until January 10, 2024, its "usage policies" included a ban on "activity that has high risk of physical harm, including", specifically, "weapons development" and "military and warfare". Its new policies prohibit "[using] our service to harm yourself or others" and to "develop or use weapons".{{cite web | url=https://theintercept.com/2024/01/12/open-ai-military-ban-chatgpt/ | title=OpenAI Quietly Deletes Ban on Using ChatGPT for "Military and Warfare" | date=January 12, 2024 | access-date=January 13, 2024 | archive-date=January 25, 2024 | archive-url=https://web.archive.org/web/20240125091034/https://theintercept.com/2024/01/12/open-ai-military-ban-chatgpt/ | url-status=live }}{{Cite news |date=January 16, 2024 |title=OpenAI Is Working With US Military on Cybersecurity Tools |url=https://www.bloomberg.com/news/articles/2024-01-16/openai-working-with-us-military-on-cybersecurity-tools-for-veterans |access-date=January 23, 2024 |work=Bloomberg.com |language=en |archive-date=January 26, 2024 |archive-url=https://web.archive.org/web/20240126131713/https://www.bloomberg.com/news/articles/2024-01-16/openai-working-with-us-military-on-cybersecurity-tools-for-veterans |url-status=live }} As one of the industry collaborators, OpenAI provides LLM to the Artificial Intelligence Cyber Challenge (AIxCC) sponsored by Defense Advanced Research Projects Agency (DARPA) and Advanced Research Projects Agency for Health to protect software critical to Americans.{{Cite web |title=OpenAI {{!}} AIxCC |url=https://aicyberchallenge.com/openai/ |website=Artificial Intelligence Cyber Challenge |access-date=March 30, 2024 |archive-date=March 30, 2024 |archive-url=https://web.archive.org/web/20240330202438/https://aicyberchallenge.com/openai/ |url-status=live }} In October 2024, The Intercept revealed that OpenAI's tools are considered "essential" for AFRICOM's mission and included in an "Exception to Fair Opportunity" contractural agreement between the United States Department of Defense and Microsoft.{{Cite web |last=Biddle |first=Sam |date=2024-10-25 |title=Pentagon Purchased OpenAI Tools for Military Operations Across Africa |url=https://theintercept.com/2024/10/25/africom-microsoft-openai-military/ |access-date=2024-10-26 |website=The Intercept |language=en-US |archive-date=October 26, 2024 |archive-url=https://web.archive.org/web/20241026211326/https://theintercept.com/2024/10/25/africom-microsoft-openai-military/ |url-status=live }} In December 2024, OpenAI said it would partner with defense-tech company Anduril to build drone defense technologies for the United States and its allies.{{Cite web |title=OpenAI's new defense contract completes its military pivot |url=https://www.technologyreview.com/2024/12/04/1107897/openais-new-defense-contract-completes-its-military-pivot |access-date=2024-12-05 |website=MIT Technology Review |language=en}}

In June 2023, a lawsuit claimed that OpenAI scraped 300 billion words online without consent and without registering as a data broker. It was filed in San Francisco, California, by sixteen anonymous plaintiffs.{{Cite web |last=Riley |first=Tonya |date=2023-06-30 |title=OpenAI lawsuit reignites privacy debate over data scraping |url=https://cyberscoop.com/openai-lawsuit-privacy-data-scraping/ |access-date=2024-11-26 |website=CyberScoop |language=en-US}} They also claimed that OpenAI and its partner as well as customer Microsoft continued to unlawfully collect and use personal data from millions of consumers worldwide to train artificial intelligence models.{{cite news |last1=Xiang |first1=Chloe |title=OpenAI and Microsoft Sued for $3 Billion Over Alleged ChatGPT 'Privacy Violations' |url=https://www.vice.com/en/article/openai-and-microsoft-sued-for-dollar3-billion-over-alleged-chatgpt-privacy-violations/ |work=Vice |date=June 29, 2023 |language=en |access-date=July 13, 2024 |archive-date=July 13, 2024 |archive-url=https://web.archive.org/web/20240713042222/https://www.vice.com/en/article/wxjxgx/openai-and-microsoft-sued-for-dollar3-billion-over-alleged-chatgpt-privacy-violations |url-status=live }}

On May 22, 2024, OpenAI entered into an agreement with News Corp to integrate news content from The Wall Street Journal, the New York Post, The Times, and The Sunday Times into its AI platform. Meanwhile, other publications like The New York Times chose to sue OpenAI and Microsoft for copyright infringement over the use of their content to train AI models.{{cite news|title=OpenAI and Wall Street Journal owner News Corp sign content deal |date=May 22, 2024 |url=https://www.theguardian.com/technology/article/2024/may/22/openai-chatgpt-news-corp-deal |archive-url=https://web.archive.org/web/20240523030813/https://www.theguardian.com/technology/article/2024/may/22/openai-chatgpt-news-corp-deal |archive-date=May 23, 2024 |access-date=May 23, 2024}} In November 2024, a coalition of Canadian news outlets, including the Toronto Star, Metroland Media, Postmedia, The Globe and Mail, The Canadian Press and CBC, sued OpenAI for using their news articles to train its software without permission.{{Cite web |title=Major Canadian news outlets sue OpenAI |url=https://www.bbc.com/news/articles/cm27247j6gno |access-date=2024-12-05 |website=www.bbc.com |date=November 29, 2024 |language=en-GB}}

{{Further|Suchir Balaji}}Suchir Balaji, a former researcher at OpenAI, was found dead in his San Francisco apartment on November 26, 2024. Independent investigations carried out by the San Francisco Police Department (SFPD) and the San Francisco Office of the Chief Medical Examiner (OCME) concluded that Balaji shot himself.

The death occurred 34 days after a New York Times interview in which he accused OpenAI of violating copyright law in developing its commercial LLMs, one of which (GPT-4) he had helped engineer. He was also a likely witness in a major copyright trial against the AI company, and was one of several of its current or former employees named in The New York Times{{'}}s court filings as potentially having documents relevant to the case. The death led to speculation and conspiracy theories suggesting he had been deliberately silenced.{{Cite web |date=2025-02-15 | first= Tomoki | last= Chien |title=Autopsy: No foul play in OpenAI whistleblower’s suicide |url=https://sfstandard.com/2025/02/14/autopsy-no-foul-play-in-openai-whistleblowers-suicide/ |access-date=2025-02-15 |website=The San Francisco Standard |language=en}}{{cite web |last=Rodgers |first=Jakob |date=2025-01-31 |title=Family of OpenAI whistleblower Suchir Balaji files lawsuit seeking San Francisco police records |url=https://www.mercurynews.com/2025/01/31/family-of-openai-whistleblower-suchir-balaji-files-lawsuit-seeking-san-francisco-police-records/ |accessdate=2025-02-22 |website=The Mercury News}} Elon Musk, Tucker Carlson, California Congressman Ro Khanna, and San Francisco Supervisor Jackie Fielder have publicly echoed Balaji's parents' skepticism and calls for an investigation, as of January, 2025.{{cite news |last=Rodgers |first=Jakob |date=January 15, 2025 |title=California Congressman Ro Khanna calls for 'full and transparent' investigation into death of OpenAI whistleblower Suchir Balaji |url=https://www.chicagotribune.com/2025/01/15/openai-whistleblower-death/ | archiveurl= https://archive.today/20250130154936/https://www.chicagotribune.com/2025/01/15/openai-whistleblower-death/ | archivedate= 2025-01-30| url-status= live |work= Chicago Tribune |agency=Bay Area News Group |access-date=January 30, 2025}}{{cite web| first= Troy | last= Wolverton | date= 2025-01-23 | title= Family, officials urge inquiry into OpenAI whistleblower's death | website= San Francisco Examiner | url=https://www.sfexaminer.com/news/technology/bay-area-officials-say-suchir-balaji-death-must-prompt-probe/article_8f1d1596-d923-11ef-8109-7f47649f523f.html | accessdate= 2025-01-24}}

In February 2025, the OCME autopsy and SFPD police reports were released. A joint letter from both agencies to the parents' legal team noted that he had purchased the firearm used two years prior to his death, and had recently searched for brain anatomy information on his computer. The letter also highlighted that his apartment's only entrance was dead-bolted from inside with no signs of forced entry.

• {{annotated link|Anthropic}}
• {{annotated link|Center for AI Safety}}
• {{annotated link|Future of Humanity Institute}}
• {{annotated link|Future of Life Institute}}
• {{annotated link|Google DeepMind}}
• {{annotated link|Machine Intelligence Research Institute}}
• {{annotated link|Model Context Protocol}}

{{notelist}}

{{Reflist}}

{{refbegin}}

• {{Cite magazine |last=Levy |first=Steven |author-link=Steven Levy |date=September 5, 2023 |title=What OpenAI Really Wants |url=https://www.wired.com/story/what-openai-really-wants |magazine=Wired |issn=1078-3148 |access-date=September 6, 2023 |archive-date=September 6, 2023 |archive-url=https://web.archive.org/web/20230906191725/https://www.wired.com/story/what-openai-really-wants/ |url-status=live }}
• {{Cite magazine |last=Duhigg |first=Charles |author-link=Charles Duhigg |date=December 1, 2023 |title=The Inside Story of Microsoft's Partnership with OpenAI |url=https://www.newyorker.com/magazine/2023/12/11/the-inside-story-of-microsofts-partnership-with-openai |magazine=The New Yorker |issn=0028-792X |access-date=January 15, 2024 |archive-date=December 22, 2023 |archive-url=https://archive.today/20231222230940/https://www.newyorker.com/magazine/2023/12/11/the-inside-story-of-microsofts-partnership-with-openai |url-status=live }}

{{refend}}

{{Commons category|OpenAI}}

• {{Official website}}

{{OpenAI navbox}}

{{Generative AI}}

{{Existential risk from artificial intelligence}}

{{Authority control}}

{{Portal bar|Companies|San Francisco Bay Area|California}}

Category:Artificial intelligence associations

Category:Artificial intelligence laboratories

Category:Non-profit organizations based in San Francisco

Category:501(c)(3) organizations

Category:Research institutes in the San Francisco Bay Area

Category:2015 establishments in California

Category:2015 in San Francisco

Category:American companies established in 2015