OpenAI vs Anthropic is a contest between two labs with the same goal and different bets. OpenAI, the maker of ChatGPT, is a consumer first company that reaches hundreds of millions of people and now sells ads on its free plans. Anthropic, the maker of Claude, was founded in 2021 by former OpenAI staff, earns most of its money from businesses and developers, and has promised that Claude will stay ad free. Their top models now cost the same per token, so the real differences are in control, business model and how each lab handles safety.
This guide compares the two companies, not just their chatbots, using each lab's own pages, Anthropic's IPO filing and OpenAI's funding announcements, checked on November 9, 2026, with model prices rechecked on November 19.
What's new (November 19, 2026): Both flagships got cheaper and match again. Anthropic's Claude Fable 5.2 (November 9) and OpenAI's GPT-6.2 Astra (November 17) both list at $8 per million input tokens and $40 per million output tokens, down from $10 and $50 for Fable 5.1 and GPT-6 Astra. OpenAI's GPT-6.2 Sol keeps the $2 and $10 Sol price. Sources: Claude pricing and OpenAI's GPT-6.2 launch post. To pick a chatbot plan only, see ChatGPT Plus vs Claude Pro vs Google AI Pro.
OpenAI vs Anthropic at a glance
| Item | OpenAI | Anthropic |
|---|---|---|
| Founded | 2015, as a nonprofit | 2021, by former OpenAI employees |
| Legal form today | OpenAI Group PBC, controlled by the nonprofit OpenAI Foundation | Anthropic, PBC, with a Long Term Benefit Trust |
| CEO | Sam Altman | Dario Amodei |
| Stated mission | Ensure AGI benefits all of humanity | Responsible development of advanced AI for the long term benefit of humanity |
| Flagship assistant | ChatGPT | Claude |
| Top API model | GPT-6.2 Astra ($8 / $40 per million tokens) | Claude Fable 5.2 ($8 / $40 per million tokens) |
| Last private valuation | $852 billion (March 31, 2026) | $965 billion (May 28, 2026) |
| Ads in the chatbot | Yes, on Free and Go plans | No, by written policy |
| Safety rulebook | Preparedness Framework | Responsible Scaling Policy |
| Stock listing | Private | Nasdaq (ANT) since November 19, 2026, IPO priced at $83 |
Why Anthropic split from OpenAI
Anthropic exists because of OpenAI. Dario Amodei was OpenAI's vice president of research and his sister Daniela Amodei was its vice president of safety and policy before they left and started Anthropic in early 2021, as TechCrunch reported when the new lab raised its first $124 million. Anthropic's IPO filing names seven cofounders in total.
The founders framed Anthropic as a safety focused lab building AI that is "reliable, interpretable, and steerable," wording that still opens Anthropic's company page. That shared origin is why the ethics debate between the two never fades.
Who controls each company
Both labs are public benefit corporations, so their boards must weigh a stated mission alongside shareholder returns. The machinery differs.
OpenAI. According to OpenAI's structure page, a recapitalization announced on October 28, 2025 turned the nonprofit into the OpenAI Foundation and the business into OpenAI Group PBC. The Foundation appoints every director of OpenAI Group and can replace them at any time. At closing it held about 26% of OpenAI Group, plus a warrant for more shares if the share price rises more than tenfold after 15 years. Microsoft held roughly 27%, and current and former employees and investors held the remaining 47%. A Foundation committee chaired by Zico Kolter oversees safety.
Anthropic. Anthropic, PBC uses a Long Term Benefit Trust instead of a controlling nonprofit. Its S-1 filing with the SEC describes five share classes: the Trust's Class T stock elects a majority of directors, and a single Class F share held by a Founder LLC of the seven cofounders carries 50.1% of the votes on general matters. Our Anthropic IPO explainer covers the share classes.
In short, a minority owner nonprofit appoints OpenAI's board, while a trust and the founders control Anthropic's.
How each lab makes money
OpenAI is consumer first. In its March 31, 2026 funding post, OpenAI said ChatGPT had more than 900 million weekly active users and over 50 million subscribers, that it was generating $2 billion in revenue per month, and that enterprise made up more than 40% of revenue. The same post said its ads pilot passed $100 million in annual recurring revenue in under six weeks. CNBC reported the same figures, which are self reported because OpenAI publishes no audited accounts.
Anthropic is business first. Anthropic's audited figures in the S-1 show revenue of $4.585 billion in 2025 and $16.286 billion in the first half of 2026. About 48% of first half revenue came through the Amazon, Google Cloud and Microsoft Azure marketplaces. In its own words from the February ad free pledge, Anthropic makes money "through enterprise contracts and paid subscriptions."
Valuation. OpenAI raised $122 billion at an $852 billion post money valuation in March. Two months later Anthropic raised $65 billion at $965 billion, per its Series H announcement, which Reuters noted put Anthropic ahead of OpenAI on paper. Amazon backs both labs: it committed up to $50 billion to OpenAI's March round and $5 billion to Anthropic's Series H. Later valuation talk is unconfirmed, so we leave it out.
Models and API prices compared
At list price the two labs now match tier for tier. Prices below are standard rates per million input and output tokens from OpenAI's API pricing page and Anthropic's models overview.
| Tier | OpenAI model | OpenAI price | Anthropic model | Anthropic price |
|---|---|---|---|---|
| Flagship | GPT-6.2 Astra | $8 / $40 | Claude Fable 5.2 | $8 / $40 |
| Upper middle | none | n/a | Claude Opus 5.5 | $4 / $20 |
| Workhorse | GPT-6.2 Sol or GPT-6.1 Sol | $2 / $10 | Claude Sonnet 5.5 | $2 / $10 |
| Balanced | GPT-6.1 Terra | $1 / $5 | none | n/a |
| Budget | GPT-6 Luna | $0.10 / $0.50 | Claude Haiku 5.5 | from $0.10 / $0.50 |
Two pricing details split them. OpenAI charges a higher rate once a prompt passes 272,000 input tokens (GPT-6 Astra goes to $20 in and $75 out), while every current Claude model has a 1 million token context window. Anthropic recommends Opus 5.5 as the default for most work and Fable 5.2 for the hardest reasoning and long agent runs; OpenAI's lineup is explained in our GPT-6 Sol vs Astra vs Luna guide. For a head to head of the two flagships, see GPT-6 Astra vs Claude Fable 5.1, and for the whole Claude range, Claude Fable vs Opus vs Mythos.
Developers use the OpenAI API or the Anthropic API; both labs also sell through Amazon Bedrock and Microsoft's cloud.
Consumer plans and ads
For everyday users the clearest difference is advertising. OpenAI began testing ads in the US on February 9, 2026. Its ads help page says ads may appear for Free and Go users, never for Plus, Pro, Business, Enterprise or Edu accounts or for users identified as under 18, and that ads don't influence answers. Free users can switch to an ads free mode with lower limits.
Anthropic took the other path on February 4, 2026, promising that Claude will stay ad free, with no sponsored links and no advertiser influence on answers.
| Plan level | ChatGPT | Claude |
|---|---|---|
| Free | Yes, ads possible | Yes, no ads |
| Budget | Go, $8 a month (ads possible) | none |
| Standard | Plus, $20 a month | Pro, $20 a month ($17 on annual billing) |
| Heavy use | Pro 100, Pro 200, Pro 500 ($100 to $500 a month) | Max ($100) and Ultra ($200) |
Prices come from OpenAI's Pro tiers page and Claude's pricing page. Both companies let consumers opt out of model training, and both exclude business workspaces from training by default. For a privacy focused comparison, see our safe AI alternative to ChatGPT guide.
Safety approach: Preparedness Framework vs RSP
Both labs tie stronger safeguards to dangerous capabilities, and both are voluntary policies, not laws.
- OpenAI's Preparedness Framework (version 2, April 2025) uses High and Critical thresholds in biology and chemistry, cybersecurity and AI self improvement. In August 2026 OpenAI concluded it could not rule out Critical cyber capability for Astra, paused some work, and told Axios it is rewriting the framework because models are nearing the thresholds it describes.
- Anthropic's Responsible Scaling Policy has been rewritten more often. Version 3.0 arrived in February 2026 with public Risk Reports and a Frontier Safety Roadmap, and the RSP page lists version 3.4, effective July 8, 2026, plus a redacted Risk Report from August.
The behavior rules differ too. OpenAI publishes a Model Spec (our explainer), while Anthropic trains Claude against a written constitution, covered in what is constitutional AI. The full side by side is in Preparedness Framework vs Responsible Scaling Policy.
Neither record is spotless: OpenAI disclosed a test model breaching Hugging Face systems (see our OpenAI sandbox escape explainer), and Axios reported Anthropic then found similar evidence for its own models.
Pros and cons of each lab
OpenAI pros
- The largest consumer reach, with ChatGPT used by more than 900 million people a week (March figure).
- The widest model range, including a $1 / $5 middle tier Anthropic lacks.
OpenAI cons
- Ads on Free and Go plans.
- No audited public financials; growth figures are self reported.
- A safety framework that OpenAI itself says is being rewritten.
Anthropic pros
- No ads on any plan, by written policy.
- 1 million token context on every current model at standard rates.
- Audited revenue and governance details on file with the SEC.
Anthropic cons
- No plan between Free and the $20 Pro tier.
- Fable 5.1 is metered on Pro and capped at half of weekly limits on Max and Ultra.
Which one should you pick?
- You want one assistant for everything and a big app ecosystem: OpenAI's ChatGPT, on Plus to avoid ads.
- You want a free assistant with no ads, or long documents: Anthropic's Claude.
- You build on the API: the list prices tie, so test both on your own prompts and compare long context and caching costs. Our LLM subscription vs API cost comparison shows how the bills add up.
FAQ
Is Anthropic part of OpenAI?
No. Anthropic is a separate company founded in 2021 by former OpenAI employees, including Dario and Daniela Amodei. The two labs are direct competitors, and Anthropic's IPO filing names OpenAI as one.
Which is bigger, OpenAI or Anthropic?
It depends on the measure. OpenAI has far more consumer users, with more than 900 million weekly ChatGPT users in March 2026. On the last private valuations, Anthropic ($965 billion in May) was ahead of OpenAI ($852 billion in March).
Is Anthropic more ethical than OpenAI?
There is no neutral score. Anthropic refuses ads and publishes frequent Risk Reports. OpenAI has a controlling nonprofit, publishes a detailed Model Spec and argues ads widen free access. Both have had safety incidents in testing.
Do OpenAI and Anthropic charge the same for their APIs?
At list price, yes for three tiers: $8 / $40 for GPT-6.2 Astra and Claude Fable 5.2, $2 / $10 for GPT-6.2 Sol and Claude Sonnet 5.5, and $0.10 / $0.50 for GPT-6 Luna and Claude Haiku 5.5. Bills still differ because of long context rates, caching and token counts.
Can I buy OpenAI or Anthropic stock?
Anthropic, yes; OpenAI, no. Anthropic priced its IPO at $83 a share on November 18, 2026, and its Class A shares have traded on the Nasdaq under the ticker ANT since November 19. OpenAI is still private. Our Anthropic IPO explainer covers the terms.
Who owns OpenAI and Anthropic?
At OpenAI's 2025 recapitalization, the OpenAI Foundation held about 26% of OpenAI Group, Microsoft about 27%, and employees and investors the rest. Anthropic is owned by its investors, employees and strategic partners such as Amazon and Google, while voting control sits with its Long Term Benefit Trust and its cofounders.



