ChatGPT was created by OpenAI, an American company based in San Francisco and founded in December 2015 by eleven people, including Sam Altman, Elon Musk, Greg Brockman and Ilya Sutskever. Launched on November 30, 2022, ChatGPT is now led by Sam Altman, CEO of OpenAI, and Greg Brockman, its president.
On the build side, the team that developed ChatGPT was led by John Schulman, an OpenAI cofounder, with Mira Murati as CTO. Below are the facts as verified on October 7, 2026, with exact dates, and the mix-ups worth retiring for good.
Who created ChatGPT, in 30 seconds
A nonprofit founded in 2015, turned capped-profit company in 2019, then public benefit corporation in 2025. ChatGPT arrived seven years after the founding, on a model that already existed.
- Creator: OpenAI, San Francisco, United States
- Launch: November 30, 2022, on GPT-3.5
- Leaders: Sam Altman (CEO), Greg Brockman (president)
- Eleven cofounders, including Elon Musk, who left in 2018
- Elon Musk did not create ChatGPT
- Microsoft does not own ChatGPT
Origin: United States, California Models in October 2026: GPT-5.6 and GPT-6
Key takeaways
- ChatGPT was created by OpenAI, an organization founded in December 2015 in San Francisco by eleven people, including Sam Altman and Elon Musk.
- ChatGPT was released on November 30, 2022, on GPT-3.5, and passed one million users in five days.
- John Schulman led the research team that built ChatGPT, with Mira Murati as CTO.
- Sam Altman runs OpenAI, with Greg Brockman as president. Elon Musk left the board in 2018.
- Since October 28, 2025, OpenAI has been a public benefit corporation (OpenAI Group PBC), controlled by the OpenAI Foundation; Microsoft then held about 27%.
Test what you think you know. Eight statements, each corrected as you go, with the exact date.
Who created ChatGPT? OpenAI and its eleven cofounders
ChatGPT is a product of OpenAI, headquartered in San Francisco, California. The organization was founded in December 2015 (publicly announced on December 11) by eleven people. The full list is rarely given, so here it is:
| Cofounder | What to know |
|---|---|
| Sam Altman | Co-chair at the founding, CEO of OpenAI since 2019 |
| Elon Musk | Co-chair at the founding, left the board in 2018, later launched xAI and its assistant Grok |
| Greg Brockman | First CTO, now OpenAI's president |
| Ilya Sutskever | Chief scientist until 2024, founder of Safe Superintelligence |
| John Schulman | Led the research team that built ChatGPT, left in 2024 |
| Andrej Karpathy | Researcher, formerly at Tesla, coined the term "vibe coding" in 2025 |
| Wojciech Zaremba | Researcher, cofounder |
| Durk Kingma | Researcher, coauthor of landmark machine learning papers |
| Trevor Blackwell | Engineer, partner at the Y Combinator startup accelerator |
| Vicki Cheung | Engineer on the founding team |
| Pamela Vagata | Engineer on the founding team |
A point most articles leave out: OpenAI was founded as a nonprofit. The founders and their backers pledged $1 billion, but according to its tax filings, about $133 million had actually been contributed by 2021. That gap explains part of what came next.
Who built the ChatGPT app?
Crediting ChatGPT to a single person is inaccurate: it is the work of hundreds of researchers and engineers. Two names do keep coming up, though. John Schulman, a cofounder, led the reinforcement learning team that developed ChatGPT, building on a method OpenAI published in January 2022 (InstructGPT): training the model with human feedback. Mira Murati, CTO of OpenAI from 2022 to 2024, oversaw the teams that shipped ChatGPT, DALL-E and Sora. She is usually the person people mean when they ask about "the woman who created ChatGPT". She left OpenAI in September 2024 and founded Thinking Machines Lab in 2025.
When was ChatGPT released?
On November 30, 2022, as a free preview. It took off instantly: more than one million users in five days, and roughly 100 million within two months according to estimates published in early 2023. By early 2026, OpenAI was claiming more than 900 million weekly users.
Two clarifications correct some very common mistakes.
- ChatGPT ran on GPT-3.5, not GPT-4. GPT-4 only arrived on March 14, 2023. So the global breakthrough did not come from a technological leap: what changed was the conversational interface, free and open to everyone.
- Seven years separate OpenAI's founding from the release of ChatGPT. Between 2015 and 2022, OpenAI released GPT-2, GPT-3 and DALL-E, but nothing that reached the general public.
Who runs OpenAI today?
As of October 7, 2026, the leadership team looks like this:
- Sam Altman, CEO.
- Greg Brockman, president, who took over responsibility for products, including ChatGPT, in July 2026 after Fidji Simo stepped away for health reasons.
- Bret Taylor, chair of the board.
- Jakub Pachocki, chief scientist, and Sarah Friar, CFO.
Sam Altman went through a crisis in November 2023: removed by the board on November 17, he was brought back on November 21, a few days later. The episode exposed the tension between the board of a nonprofit and a business that had grown enormous. The 2025 restructuring partly stems from it.
As for Elon Musk, he left the board in 2018 and has had no role at OpenAI since. He sued the company in 2024; on May 18, 2026, a California jury rejected his claim, finding it had been filed too late.
OpenAI history: from nonprofit to public benefit corporation
The story reads as three legal structures, and that is the clearest way to understand it.
2015 to 2019: the nonprofit
OpenAI is incorporated in Delaware with an open research mission, funded by donations and funding pledges. The model quickly falls short: training these systems costs sums that no philanthropic structure can raise.
2019: the capped-profit structure
In March 2019, OpenAI creates a for-profit subsidiary whose returns are capped at one hundred times the investment. The setup opens the door to Microsoft, which invests $1 billion in July 2019, then several billion more in January 2023, and provides the computing infrastructure needed for the next models.
Since October 28, 2025: the public benefit corporation
The for-profit arm becomes OpenAI Group PBC, a public benefit corporation controlled by the OpenAI Foundation. The split published that day: about 26% for the foundation, 27% for Microsoft, 47% for employees and other investors. In 2026, OpenAI closed a $122 billion round at an $852 billion valuation, with SoftBank, Amazon and Nvidia among the investors, so the ownership split has shifted again since.
OpenAI and ChatGPT timeline, from 2015 to GPT-6
| Date | Event |
|---|---|
| December 2015 | OpenAI founded as a nonprofit |
| February 2018 | Elon Musk leaves the board |
| February 2019 | GPT-2, released in stages out of caution |
| March and July 2019 | Capped-profit subsidiary, then $1 billion from Microsoft |
| 2020 | GPT-3 and its first commercial API |
| January 2021 | DALL-E, image generation from text |
| January 2022 | InstructGPT, training with human feedback |
| November 30, 2022 | ChatGPT launches, on GPT-3.5 |
| March 14, 2023 | GPT-4 |
| November 17 to 21, 2023 | Sam Altman removed, then reinstated |
| May 2024 | GPT-4o; Ilya Sutskever leaves |
| September to December 2024 | o1, the first reasoning model; Mira Murati leaves; public launch of Sora |
| August 7, 2025 | GPT-5 |
| October 28, 2025 | OpenAI becomes OpenAI Group PBC |
| December 2025 | GPT-5.2 |
| March 24, 2026 | Sora shutdown announced |
| May 18, 2026 | A jury rejects Elon Musk's lawsuit |
| July 9, 2026 | GPT-5.6 (Sol, Terra, Luna) and launch of the ChatGPT Work agent |
| September 2026 | GPT-6 Astra, then GPT-6 Sol and GPT-6 Luna |
In October 2026, free ChatGPT users chat with GPT-5.6 Luna, Plus subscribers with GPT-5.6 Sol, and Pro plans give access to GPT-6 Pro, powered by GPT-6 Astra. Remember the logic rather than the version numbers: the release pace has sped up sharply since 2024.
GPT vs ChatGPT: what's the difference?
GPT is the family of models: GPT-2, GPT-3, GPT-4, then GPT-5 and its variants up to GPT-6. These are language models, available in particular through an API.
ChatGPT is the app you use to talk to those models: a product, with its own interface, subscriptions and features.
The distinction has practical consequences. In ChatGPT, the model that answers depends on your plan and your settings. When you build GPT into software, you choose the model explicitly and pay per use. To get the most out of the app, the method is in my guide to ChatGPT prompts.
The mistakes I see most often
Five claims circulate widely and are wrong. The quiz above covers them; here they are in detail.
"Elon Musk created ChatGPT." He is one of the eleven cofounders of OpenAI, but he left the board in 2018, four years before the app came out.
"Microsoft owns ChatGPT." Microsoft is a major shareholder and OpenAI's long-standing cloud partner, but the company is controlled by the OpenAI Foundation.
"ChatGPT launched with GPT-4." No, with GPT-3.5. This mistake suggests the success came from a technological leap, when it came from access.
"Sam Altman was fired from OpenAI." He was removed on November 17, 2023 and reinstated on the 21st. Telling the first half without the second paints a false picture.
"OpenAI is a startup like any other." When it was created, it was a nonprofit, almost the opposite. The word hides the trajectory, which is precisely what makes the story interesting.
How to check this kind of information
Don't ask a model for a date. It will produce one, plausible and sometimes wrong, with the same confidence either way. Dates, names and figures should be checked against the primary source. Separate what is confirmed from what is merely repeated, as I do for the origin of the name Claude. And date what you write: a list of executives or a model lineup is only true as of a given date.
Why November 2022 changed everything
The model behind ChatGPT at launch was not new: GPT-3 had been available to developers since 2020. What was missing was not capability, it was access.
Three choices tipped the balance. Free access, which removes the purchase decision. Conversation, which replaces a programming interface with something everyone already knows how to use. And immediate access for everyone, with no waitlist.
I stress this point because it has practical consequences. In the projects I work on, the question that decides success is almost never "which model" but "how do people get to it". An excellent tool behind a painful interface goes unused. That is the common thread of my AI automation work. And because ChatGPT has become a gateway to information, showing up in its answers has become a discipline of its own: I cover it in my article on ChatGPT SEO.
OpenAI and Anthropic: a shared origin
In late 2020, several OpenAI researchers and executives, including Dario Amodei, then VP of research, left the organization. In January 2021, they founded Anthropic, the company behind Claude, set up from day one as a public benefit corporation. The Anthropic side of the story is in my article on the origin of Claude AI.
The useful lesson when you compare these tools: they do not come from different worlds. The teams know each other, techniques circulate, and differences in behavior reflect different trade-offs made on a shared foundation. To choose between the two day to day, I compared Claude vs ChatGPT use case by use case.
Frequently asked questions
Who created ChatGPT?
ChatGPT was created by OpenAI, an American company founded in December 2015 in San Francisco by eleven people: Sam Altman, Elon Musk, Greg Brockman, Ilya Sutskever, Wojciech Zaremba, John Schulman, Andrej Karpathy, Durk Kingma, Trevor Blackwell, Vicki Cheung and Pamela Vagata. The app launched on November 30, 2022.
Who created ChatGPT, and in which country?
ChatGPT was created in the United States by OpenAI, headquartered in San Francisco, California. The company is run by Sam Altman. It is an American company, although it also has an office in Paris.
Which woman is credited with creating ChatGPT?
The question usually refers to Mira Murati, OpenAI's CTO from 2022 to 2024, who oversaw the teams behind ChatGPT. The research team that built it was led by John Schulman, an OpenAI cofounder. ChatGPT remains the collective work of hundreds of people.
Who owns ChatGPT?
ChatGPT belongs to OpenAI Group PBC, a public benefit corporation controlled by the OpenAI Foundation. At the October 28, 2025 restructuring, the foundation held about 26% of the equity, Microsoft 27%, and employees and other investors 47%. Those stakes have shifted with the 2026 funding rounds.
When was ChatGPT released?
On November 30, 2022, as a free preview. It ran on GPT-3.5, not GPT-4, which arrived on March 14, 2023. ChatGPT passed one million users in five days.
Did Elon Musk create ChatGPT?
No. Elon Musk is one of OpenAI's eleven cofounders from 2015, but he left the board in 2018, four years before ChatGPT came out. His lawsuit against OpenAI was rejected by a California jury on May 18, 2026.
Who runs OpenAI today?
Sam Altman, as CEO, with Greg Brockman as president and head of products since July 2026. Bret Taylor chairs the board. Sam Altman was briefly removed in November 2023, then brought back a few days later.
What is the difference between GPT and ChatGPT?
GPT is OpenAI's family of language models, from GPT-2 in 2019 to GPT-6 in 2026. ChatGPT is the app that lets you talk to those models. In the app, the model depends on your plan; through the API, you choose it and pay per use.
Choosing the right AI tools
I train teams and build the automations that put these models to work on your real files, whatever the provider.
Written by
Anthony Courtin
Reader reviews
Did this article help you? Say so, it helps the next readers.
What readers say
Be the first to leave a review.
