Claude Opus 5: benchmarks, price and rank
Claude Opus 5 from Anthropic is 9th of 123 rated AI models on Spezia's leaderboard, strongest on Humanity's Last Exam and BrowseComp. It costs $5 in, $25 out per million tokens.
- 9th of 123 rated models
- Rating 1,844
- Rated on 16 benchmarks
- Closed
- Released Jul 24, 2026
- 1 million tokens of context
Where it is strongest
- 2nd of 74 on Humanity's Last Exam
- 4th of 56 on BrowseComp
- 3rd of 6 on ARC-AGI-3
- 5th of 107 on LMArena text
- 9th of 98 on LMArena WebDev
Where it falls behind
- 12th of 19 on HealthBench Professional
- 14th of 27 on DeepSWE 1.1
Price: $5 in, $25 out per million tokens.
Its headline results
The benchmarks that count most in the rating, with its place among the models measured the same way, and where each number came from.
- APEX-Agents65.8%
Long tasks written by investment banking analysts, management consultants and corporate lawyers from their own work, done across files and apps in a realistic work setup.
11th of 56. Independent result from Epoch AI.
- Terminal-Bench 4.051.8%
Hard jobs done in a real computer terminal: building and fixing software, setting up systems, and science and engineering problems.
Independent result from LLM Stats.
- BrowseComp90.8%
Questions whose answers are hidden deep on the web and take many searches and pages to track down.
4th of 56. The maker's own number, from Anthropic, via LLM Stats.
- FrontierSWE52.0%
Open-ended technical projects that take hours to tens of hours, from making systems faster to building large programs and applied machine learning research. Each run gets a 20-hour budget.
8th of 21. Independent result from Epoch AI.
- Humanity's Last Exam64.7%
About 2,500 questions written by experts in more than a hundred fields to be as hard as they could make them, many needing a diagram or image.
2nd of 74. The maker's own number, from Anthropic, via LLM Stats.
- ARC-AGI-330.2%
Interactive puzzle games the model has never seen, with no instructions. It has to explore, work out the goal and the rules, and win.
3rd of 6. The maker's own number, from Anthropic, via LLM Stats.
- Vending-Bench 2$11,182
The model runs a simulated vending machine business for a year on its own: finding suppliers, ordering stock, setting prices and paying the daily fees.
6th of 60. Independent result from Epoch AI.
- CursorBench46.6%
Ambiguous coding tasks that span many files, taken from real sessions in the Cursor code editor.
10th of 20. Independent result from Epoch AI.
- DeepSWE 1.168.8%
Software engineering tasks run in the same simple agent setup for every model.
14th of 27. The maker's own number, from Anthropic, via LLM Stats.
- SimpleBench80.6%
Multiple-choice puzzles about everyday situations, space, time and people, written to trip up models that pattern-match instead of think. Ordinary adults do well on them.
11th of 68. Independent result from Epoch AI.
11 more results in its full profile.
Every result with its source, what it costs per task, what jumped with its release, side by side with any other model, and answers to your questions about when to use it.
Models ranked near it
How the rating works, and every benchmark it uses: the model leaderboard.