Lists / llm apis / Head to head
OpenAI vs Anthropic (2026): What ChatGPT, Claude & Gemini Say
OpenAI was named in 50 of 50 AI answers, Anthropic in 46 of 50. See which of 10 models lean which way, and when each API fits, from one dated panel.
AI models name OpenAI more often, but only just. In the 2026-09 panel, OpenAI was named in 50 of 50 answers and Anthropic in 46 of 50. The real gap is order. OpenAI came first in 33 of 50 answers, Anthropic in 14 of 50. Both reach almost every shortlist, and OpenAI usually heads it.
This page counts which LLM API ten AI models name when a developer asks what to build on. It does not rate either API.
TL;DR
- OpenAI appears in every recorded answer from all ten models. Anthropic misses only in two OpenAI-built models.
- Order separates the two more than presence does. OpenAI is usually listed first and Anthropic usually second.
- The Claude models name OpenAI in every answer. The only models that ever leave either vendor out are OpenAI’s own.
- The answers cast OpenAI as the default platform and Anthropic as the pick for coding and agent work. Choose on that fit, then benchmark both on your own workload, as several answers advise.
How often do AI models recommend OpenAI and Anthropic?
OpenAI is named in 50 of 50 recorded answers and Anthropic in 46 of 50, so both sit on nearly every shortlist the models produce. Position is where they part. OpenAI is named first in 33 answers and Anthropic in 14, and OpenAI’s average position is 1.52 against Anthropic’s 1.98. Google sits between them in the category, named in every answer but first in only 2.
| Vendor | Named | Answer share | Named first | First share | Average position | Category rank |
|---|---|---|---|---|---|---|
| OpenAI | 50/50 | OpenAI 100% | 33/50 | OpenAI 66% | 1.52 | 1 of 15 |
| Google (context) | 50/50 | Google 100% | 2/50 | Google 4% | 2.88 | 2 of 15 |
| Anthropic | 46/50 | Anthropic 92% | 14/50 | Anthropic 28% | 1.98 | 3 of 15 |
Answer share is the share of recorded answers that named the vendor. Named first is the share of answers where it appeared before any other tracked product.
The named-versus-first gap tells the story the totals hide. For OpenAI it is 34 points. For Anthropic it is 64 points. Anthropic turns up nearly as often, but it usually turns up after OpenAI.
Average position adds a second reading. When Anthropic is named, it lands close to the top of the answer. Its 1.98 sits well ahead of Google’s 2.88, even though Google appears in more answers. Read together, the counts suggest a front pair with a settled order, and Google as a name that comes later. The full category, all 15 vendors, is in the LLM API index.
Which models prefer OpenAI, and which prefer Anthropic?
No model in the panel names Anthropic more often than OpenAI. OpenAI appears in all five answers from each of the ten models, so no model names any API more often. Anthropic matches that in eight models. It falls short only in two of OpenAI’s own: GPT-5.6 Sol named it in 2 of 5 answers and ChatGPT in 4 of 5.
| Model | Maker | OpenAI named | Anthropic named |
|---|---|---|---|
| GPT-5.6 Sol | OpenAI | 5/5 | 2/5 |
| ChatGPT | OpenAI | 5/5 | 4/5 |
| GPT-5.6 Luna | OpenAI | 5/5 | 5/5 |
| Claude Opus 5 | Anthropic | 5/5 | 5/5 |
| Claude | Anthropic | 5/5 | 5/5 |
| Claude Fable 5 | Anthropic | 5/5 | 5/5 |
| Gemini | 5/5 | 5/5 | |
| Gemini 3.5 Flash | 5/5 | 5/5 | |
| Perplexity | Perplexity | 5/5 | 5/5 |
| Sonar Reasoning Pro | Perplexity | 5/5 | 5/5 |
The sharpest disagreement sits inside one maker. GPT-5.6 Luna named Anthropic in every answer. GPT-5.6 Sol named it in two. Same vendor, same five questions, different shortlists. Across the OpenAI family’s 15 answers, Anthropic’s share is Anthropic 73.3%. In the Anthropic, Google and Perplexity families it is Anthropic 100%.
The mirror case does not happen. Claude Opus 5, Claude and Claude Fable 5 each name OpenAI in all five answers. The Claude models do not leave their maker’s rival out.
The four models made by neither vendor are the cleanest read. Both Gemini models and both Perplexity models name OpenAI and Anthropic in every answer. For a buyer, that is the useful signal: outside the two vendors’ own models, the pair is treated as equally present, and the difference is only in which one gets listed first.
The heading asks about preference. The data measures presence. Which vendor each model lists first, answer by answer, is in the full record at the LLM API index.
What do the answers say about each?
The recorded answers tend to name OpenAI as the default platform and Anthropic as the choice for coding and agent work. These five lines are quoted from the answers as recorded. They are the models’ claims, not findings of this page.
- Perplexity (Sonar Pro): “OpenAI is usually the safest pick for ecosystem depth and tooling, while Anthropic is often preferred for top-tier model quality and agentic reliability.”
- GPT-5.6 Luna: “Start with OpenAI’s API unless you have a specific reason not to.”
- Gemini 3.5 Flash, in a section heading: “Best for Logic, Coding, & Agentic Workflows: Anthropic (Claude API)”
- Perplexity (Sonar Pro), in a different answer: “I’d recommend starting with Anthropic’s Claude API as the default choice”
- Claude Opus 5, before giving its view: “I’m made by Anthropic, which sells one of the APIs you’d be choosing between (Claude).”
The last line matters for reading the split above. Claude Opus 5 disclosed its maker and still named OpenAI in all five answers.
How do OpenAI and Anthropic differ?
The captured record separates them on strategy and customer, not on price. No pricing model is recorded for either vendor in this edition, and none of the captured pages lists API prices. Check each vendor’s own pricing page for current rates.
How each earns
The captured coverage describes revenue, not rate cards. Euronews reports that only a small proportion of ChatGPT users pay a subscription fee. The same report says Claude Code, Anthropic’s assistant for developers, has become one of the company’s most popular products. PitchBook analyst Harrison Rolfes, interviewed by Morningstar, says investors must believe Anthropic can convert surging consumption into durable recognised revenue. Rolfes also says OpenAI shares revenue with Microsoft.
Who each is for
Anthropic’s record points at developers and business buyers. Euronews reports that Anthropic pursued a narrower strategy, focusing on AI tools for software developers and businesses. PitchBook’s profile, reproduced by Morningstar, describes Anthropic as the developer of a large language model platform designed for business-critical use cases. CMC Markets reports that Claude models gained traction with enterprise users for their safety features and long-context capabilities.
OpenAI’s record is broader. Euronews reports that OpenAI expanded into areas ranging from video generation to web browsers. The same report says OpenAI has reportedly scaled back work on video generation. CMC Markets reports that Sam Altman refocused OpenAI on competing with Anthropic for enterprise demand and defending ChatGPT’s position.
Structure and backing
OpenAI began as a non-profit. CMC Markets reports it was founded as a non-profit AI research lab. It later shifted to a “capped-profit” structure to attract capital and scale development. The same page says Microsoft owns a diluted stake in the for-profit part of the organisation.
Anthropic grew out of OpenAI. Euronews reports it was founded by former OpenAI executives. Rolfes says Anthropic has diversified its compute across Amazon’s AWS, Google, Microsoft, SpaceX and other providers.
What the models name each for
The answers split them by job. OpenAI is named as the general default and the widest ecosystem. Anthropic is named for coding, agents and long, careful work. That reading comes from the quoted answers above, and it is the models’ description, not a measured capability.
When should you pick OpenAI?
Pick OpenAI if you want the API that every model in this panel names in every answer, and usually names first. On the measured numbers it leads Anthropic on named count, named-first count and average position. It trails Anthropic in no model.
- Named in all 50 answers, by all ten models across four makers.
- Placed first in 33 of 50 answers (OpenAI 66%), more than any other tracked vendor.
- OpenAI’s own documentation feeds the answers. developers.openai.com was cited 13 times and platform.openai.com 8 times. No Anthropic-owned host is among the five most-cited vendor hosts.
- Worth weighing if your buyers include defence agencies. SoftwareSeni reports that OpenAI signed its own Pentagon deal within about a day of Anthropic’s contract collapsing.
When should you pick Anthropic?
Pick Anthropic if your product centres on coding or agent work and you want an API that eight of the ten models name in every answer. It leads OpenAI on no headline count. Where it comes out level is presence: outside two OpenAI models, it is named as often as OpenAI.
- Present in all five answers from eight models, including both Gemini models and both Perplexity models.
- Listed first in 14 of 50 answers (Anthropic 28%), the second-highest count in the category after OpenAI.
- Close to the top when named, with an average position of 1.98.
- Worth weighing if your customers care about barred uses. SoftwareSeni reports that Anthropic refused to let its models be used for fully autonomous weapons and mass domestic surveillance.
How this sits against the OpenAI vs Anthropic guides
None of the four pages captured from the Australian results for “OpenAI vs Anthropic” compares the two APIs. They cover the companies as investments and as defence suppliers.
Morningstar Australia runs part two of Tom Lauricella’s Q&A with PitchBook analyst Harrison Rolfes on Anthropic’s challenges and worth. It covers compute, energy, trust and valuation ahead of the IPO. The page carries a general-advice disclaimer.
SoftwareSeni publishes James A. Wondrasek’s account of why Anthropic refused the Pentagon deal and OpenAI accepted. Its advice to buyers is to price vendor posture into supply-chain and reputational due diligence. SoftwareSeni sells managed team extensions and custom software. The page carries banners for Cybermatika, a product listed in SoftwareSeni’s own menu.
Yahoo Finance Australia carries a Euronews piece by Doloresz Katanich on five things to know before Anthropic’s Wall Street debut.
CMC Markets’ OPTO section publishes a piece by Aureon on what to expect from OpenAI’s and Anthropic’s IPOs. CMC Markets describes itself as an execution-only service provider. Its disclaimer says nothing in the material is financial, investment or other advice.
A PCMag security report also ranks, but it could not be retrieved and is not described here.
Those pages judge the companies as businesses. This page counts what AI answers name when a developer asks which API to build on. None of them reports that.
How the sample was built
Each answer comes from a fixed panel: 10 models x 5 fixed prompts = 50 recorded answers, one answer per model and question. The five questions, verbatim:
- What is the best LLM API to build a product on for a developer? Name specific products.
- Which LLM API to build a product on would you recommend to a developer in 2026?
- Compare the top LLM API to build a product on options right now.
- I’m a developer and I need a LLM API to build a product on. What should I use and why?
- Best LLM API to build a product on for an AI startup balancing cost and quality?
The ten models come from four makers. OpenAI’s three (GPT-5.6 Sol, ChatGPT, GPT-5.6 Luna) gave 15 answers. Anthropic’s three (Claude Opus 5, Claude, Claude Fable 5) gave 15. Google’s two (Gemini, Gemini 3.5 Flash) gave 10, and Perplexity’s two (Perplexity, Sonar Reasoning Pro) gave 10. In this edition ChatGPT is GPT-5.6 Terra, Claude is Claude Sonnet 5, Gemini is Gemini 3.6 Flash and Perplexity is Sonar Pro. Every model ran with web search on.
An answer counts for OpenAI when it names OpenAI or its GPT models, and for Anthropic when it names Anthropic or Claude. The full method sets out what is counted and how.
What these counts cannot tell you
Visibility is in scope and product quality is not. A count says how often the models named a vendor. It says nothing about uptime, support, price fairness or fit with a particular stack.
Being named differs from being recommended. An answer can list a vendor only to warn against it.
Six of the ten models are made by the two vendors on this page. The split shows where that matters, but the panel cannot say why a model leaves a vendor out.
Each model gave one answer per question, in English, in one dated edition. Answers from the API can differ from the same model’s consumer chat app. Counting is string-based, so a passing mention counts the same as a recommendation. The models searched the web, so their answers partly reflect the pages they retrieved.
The panel is this publication’s own fixed selection, not an independent award or a universal ranking. No vendor pays to appear, be reordered or be removed.
Frequently asked questions
Which is bigger, OpenAI or Anthropic?
In AI answers about LLM APIs, OpenAI is the more visible. It was named in 50 of 50 answers against Anthropic’s 46 of 50, and first in 33 against 14. The panel does not measure company size. CMC Markets heads its OpenAI section “from Upstart to Incumbent”. It heads its Anthropic section “Conscientious Challenger”. For valuation and revenue, the finance coverage described above is the place to look.
Will Anthropic surpass OpenAI?
The panel cannot forecast. In this edition OpenAI leads on named count, named-first count and average position. Euronews reports that Anthropic was widely seen as trailing OpenAI until recently. Editions are monthly, so each new edition shows whether Anthropic’s named-first count closes on OpenAI’s.
Are Anthropic and OpenAI the same company?
No. They are separate companies with separate APIs. Euronews reports that Anthropic was founded by former OpenAI executives. The panel tracks them as two vendors with separate counts and separate records.
Do Claude models favour Anthropic in their answers?
Not in these counts. Claude Opus 5, Claude and Claude Fable 5 name Anthropic in every answer, but they name OpenAI in every answer too. All of Anthropic’s missing mentions come from GPT-5.6 Sol and ChatGPT, not from the Claude models.
Is being named in an AI answer the same as being recommended?
No. A name counts whenever an answer mentions the vendor, including in a warning or a passing comparison. The count shows presence on the shortlist, and the recorded answers show the reason each model gave. The method explains what is counted.