Compare / Anthropic vs OpenAI

AI labs compared · October 2026

Anthropic vs OpenAI: which AI lab is ahead?

Each lab's best model on sale goes head to head on every kind of work we measure, from our own benchmarks and licensed sources. 43 models in all; a lab is only as good as the model you would actually pick from it.

16 models on sale with our resultsAnthropic Best: Claude Opus 5.5 · newest: Claude Haiku 5.5
27 models on sale with our resultsOpenAI Best: GPT-6.1 Sol · newest: GPT-6.1 Sol
Anthropic wins 41 too close to callOpenAI wins 5

Kinds of work led by each lab's best model there, of the 10 both are measured on.

Results as of 8 October 2026 (catalogue release 2026-10-08-4199ce0620fe). Models on sale means listed on OpenRouter today.

Anthropic or OpenAI? The short answer

OpenAI is ahead: its best models beat Anthropic's in 5 of the 10 kinds of work both are measured on; Anthropic leads in 4, and 1 is a tie.

  • Pick Anthropic for agents and tool use, reasoning and knowledge, writing and sticking to the facts.
  • Pick OpenAI for product listings, decks from an analysis, user surveys, marketing planning and speed under pressure.
  • On our overall leaderboard, Claude Opus 5.5 is #1 (85 of 100) and GPT-6.1 Sol is #2 (81).
  • Prices start at $0.20 per million tokens for Anthropic (Claude Haiku 5.5) and $0.060 for OpenAI (gpt-oss-20b).

Head to head

85Claude Opus 5.5
Best Spring Prompt overall score
81GPT-6.1 Sol
57.6Claude Opus 5.5
Best Intelligence Index
52.7GPT-6 Astra
$8.00Claude Opus 5.5
Flagship price, per million tokens
$4.00GPT-6.1 Sol
$0.20Claude Haiku 5.5
Cheapest model, per million tokens
$0.060gpt-oss-20b
16
Models on sale with our results
27

Each line names the model behind the number. Prices are blended, three input tokens to one output, at the developer's list price where we have read it; see every model's price. Intelligence Index from Artificial Analysis; overall score from our leaderboard.

Which lab is better for what

For each kind of work, each lab's best model there and its rank among every model on sale. Further right is better; the better rank is in green.

Product listingsTurning a sparse product feed and photos into listings that can go live 6 of 20Claude Opus 5.5 1 of 20GPT-6 Astra
Decks from an analysisTurning a finished analysis into a deck you could present as it is 4 of 19Claude Sonnet 5.5 1 of 19GPT-6 Astra
User surveysPlanning a user survey and reading its results without being misled 2 of 8Claude Opus 5.5 1 of 8GPT-6.1 Sol
Marketing planningPlanning a year of ad spend without overspending 4 of 19Claude Sonnet 5.5 1 of 19GPT-6 Astra
Agents and tool useMulti-step tasks with tools: support desks, coding agents, function calls 1 of 164Claude Sonnet 5.5 3 of 164GPT-6.1 Sol
Professional workReal tasks from banking, consulting and law, business documents and freelance projects 1 of 51Claude Sonnet 5.5 1 of 51GPT-5.6 Sol
Reasoning and knowledgeHard questions across science, maths and general knowledge 1 of 198Claude Opus 5.5 3 of 198GPT-6 Astra
WritingWhat people prefer in blind comparisons, and judged writing quality 1 of 188Claude Fable 5 14 of 188GPT-5.5
Sticking to the factsSummarising without inventing things, and factual answers 2 of 152Claude Opus 5.5 18 of 152GPT-6.1 Sol
Speed under pressureGood decisions against a real clock (fast chess) 15 of 22Claude Haiku 5.5 10 of 22GPT-6 Sol

Every model, by intelligence and price

Each dot is a model on sale with an Intelligence Index from Artificial Analysis. Up is smarter; left is cheaper. Hover or tap a dot for its name.

AnthropicOpenAI

020406080 $0.01$0.1$1$10$100 Price per million tokens, blended (log scale) Intelligence Index Claude Opus 5.5Claude Fable 5.1Claude Sonnet 5.5Claude Opus 5GPT-6.1 SolGPT-6 Astra

On our overall leaderboard

The 9 of their models we rank overall: one score out of 100 across everything we publish (how the score works). Only models with enough of our own results are ranked.

  1. #1Claude Opus 5.585
  2. #2GPT-6.1 Sol81
  3. #3GPT-6 Astra79
  4. #4GPT-6 Sol69
  5. #5Claude Fable 5.166
  6. #7Claude Sonnet 5.564
  7. #13Claude Haiku 5.544
  8. #15GPT-6 Luna40
  9. #22Claude Haiku 4.57

Benchmark by benchmark

The headline result of every benchmark both labs have models on: each lab's best model and its rank among the models on sale. Green marks the better rank.

MeasureAnthropicOpenAI
Agents and tool use
τ²-bench · Task success (pass^1) · tau2 v1.0.1, retail% of tasks, higher is better 79.6%Claude Opus 4.5 · 3 of 681.6%GPT-5.2 · 2 of 6
τ²-bench · Task success (pass^1) · tau2 v1.0.1, airline% of tasks, higher is better 84.0%Claude Opus 4.5 · 1 of 683.0%GPT-5.2 · 2 of 6
τ²-bench · Task success (pass^1) · tau2 v1.0.1, telecom% of tasks, higher is better 92.3%Claude Opus 4.5 · 2 of 689.7%GPT-5.2 · 4 of 6
Berkeley Function Calling Leaderboard (BFCL) V4 · Overall accuracy · BFCL% correct, higher is better 77.5%Claude Opus 4.5 · 1 of 3863.1%o3 · 5 of 38
OpenHands Index · Average% resolved, higher is better 81.0%Claude Fable 5 · 1 of 3365.9%GPT-5.5 · 5 of 33
Microsoft STATE-Bench · Task success (pass@1) · STATE-Bench v0.7% of tasks, higher is better 53.4%Claude Opus 4.7 · 2 of 455.7%GPT-5.4 · 1 of 4
Artificial Analysis · Terminal-Bench 4.0% of tasks, higher is better 63.6%Claude Sonnet 5.5 · 1 of 10559.6%GPT-6 Astra · 2 of 105
Artificial Analysis · Terminal-Bench 2.1% of tasks, higher is better 91.4%Claude Fable 5.1 · 1 of 11589.9%GPT-6 Astra · 2 of 115
Artificial Analysis · τ-bench banking% of tasks, higher is better 47.2%Claude Fable 5.1 · 7 of 11344.3%GPT-5.6 Sol · 12 of 113
Vending-Bench 2 · Money after a yearUS dollars, higher is better $11,182Claude Opus 5 · 3 of 56$15,515GPT-6 Astra · 1 of 56
Decks from an analysis
DeckBench · Deck ratingrating, higher is better 1,352Claude Sonnet 5.5 · 4 of 191,477GPT-6 Astra · 1 of 19
Marketing planning
ROASBench · Overall scorescore out of 100, higher is better 52.0Claude Sonnet 5.5 · 4 of 1956.6GPT-6 Astra · 1 of 19
Product listings
CatalogBench · Reliably publish-ready% of products, higher is better 55.4%Claude Opus 5.5 · 5 of 2075.0%GPT-6.1 Sol · 1 of 20
CatalogBench · Reliably publish-ready · sales brief% of products, higher is better 21.4%Claude Opus 5.5 · 7 of 2069.6%GPT-6 Astra · 1 of 20
Professional work
APEX-Agents · Tasks passed% of tasks, higher is better 75.5%Claude Sonnet 5.5 · 1 of 3764.7%GPT-6 Astra · 7 of 37
GDP.pdf · Professional document tasks% of rubric, higher is better 30.6%Claude Opus 5.5 · 2 of 3130.7%GPT-5.6 Sol · 1 of 31
Remote Labor Index · Projects done to client standard% of projects, higher is better 17.9%Claude Fable 5.1 · 2 of 1220.8%GPT-6 Astra · 1 of 12
Reasoning and knowledge
Artificial Analysis · Artificial Analysis Intelligence Indexindex score, higher is better 57.6Claude Opus 5.5 · 1 of 19852.7GPT-6 Astra · 4 of 198
Artificial Analysis · Humanity's Last Exam% of questions, higher is better 61.4%Claude Opus 5.5 · 1 of 19754.7%GPT-6 Astra · 6 of 197
Artificial Analysis · GPQA Diamond% of questions, higher is better 93.7%Claude Fable 5.1 · 8 of 18796.3%GPT-6 Astra · 1 of 187
Speed under pressure
BulletBench · Ladder Elo · Lightning 10+1ladder Elo, higher is better 33Claude Sonnet 5.5 · 12 of 14571GPT-5.4 mini · 4 of 14
BulletBench · Ladder Elo · Bullet 60sladder Elo, higher is better 445Claude Haiku 4.5 · 10 of 20445GPT-6 Sol · 10 of 20
Sticking to the facts
Vectara Hallucination Leaderboard · Hallucination rate · Vectara% of summaries, lower is better 9.8%Claude Haiku 4.5 · 43 of 793.1%GPT-5.4 nano · 1 of 79
Arena (formerly LMArena) · Overall · Text factualityArena rating, higher is better 1,507Claude Opus 5.5 · 1 of 1131,486GPT-5.5 · 5 of 113
SimpleQA Verified (Epoch AI) · Correct answers% of questions, higher is better 72.2%Claude Opus 5.5 · 4 of 7375.6%GPT-6 Astra · 1 of 73
User surveys
SurveyBench · Analysis ratingrating, higher is better 1,646Claude Opus 5.5 · 2 of 81,650GPT-6.1 Sol · 1 of 8
Writing
Arena (formerly LMArena) · Overall · TextArena rating, higher is better 1,505Claude Opus 4.6 · 1 of 1661,484GPT-5.6 Sol · 14 of 166
UGI Leaderboard · Writing score · UGIscore out of 100, higher is better 74.2Claude Fable 5 · 3 of 13869.7GPT-5.5 · 11 of 138

Their models, one against one

Quick answers

Is Anthropic or OpenAI better?

OpenAI is ahead: its best models beat Anthropic's in 5 of the 10 kinds of work both are measured on; Anthropic leads in 4, and 1 is a tie.

What is Anthropic's best model?

Claude Opus 5.5: #1 on the Spring Prompt overall leaderboard and 57.6 on the Artificial Analysis Intelligence Index.

What is OpenAI's best model?

GPT-6.1 Sol: #2 on the Spring Prompt overall leaderboard and 51.8 on the Artificial Analysis Intelligence Index.

Which is better for agents and tool use, Anthropic or OpenAI?

Anthropic: its best model there, Claude Sonnet 5.5, ranks 1 of 164 models on sale, against 3 for OpenAI's GPT-6.1 Sol.

Which is better for reasoning and knowledge, Anthropic or OpenAI?

Anthropic: its best model there, Claude Opus 5.5, ranks 1 of 198 models on sale, against 3 for OpenAI's GPT-6 Astra.

Which is better for writing, Anthropic or OpenAI?

Anthropic: its best model there, Claude Fable 5, ranks 1 of 188 models on sale, against 14 for OpenAI's GPT-5.5.

Which is cheaper, Anthropic or OpenAI?

OpenAI's cheapest model with results, gpt-oss-20b, costs $0.060 per million tokens (three input to one output), against $0.20 for Claude Haiku 5.5 from Anthropic.

Other labs compared

Which model is best for your product?

Benchmarks aren't your data. We test the models from Anthropic, OpenAI and the rest on your own tasks, with the cost of every answer.

Talk to usSee the leaderboard