Claude Opus 4.7: benchmarks, price and rank
Claude Opus 4.7 from Anthropic is 30th of 123 rated AI models on Spezia's leaderboard, strongest on CharXiv Reasoning and LMArena text. It costs $5 in, $25 out per million tokens.
- 30th of 123 rated models
- Rating 1,496
- Rated on 15 benchmarks
- Closed
- Released Apr 16, 2026
- 1 million tokens of context
Where it is strongest
- 2nd of 45 on CharXiv Reasoning
- 11th of 107 on LMArena text
Where it falls behind
- 11th of 15 on CyberGym
- 33rd of 56 on APEX-Agents
- 35th of 68 on SimpleBench
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-Agents49.2%
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.
33rd of 56. Independent result from Epoch AI.
- BrowseComp79.3%
Questions whose answers are hidden deep on the web and take many searches and pages to track down.
21st of 56. The maker's own number, from Anthropic, via LLM Stats.
- MCP Atlas77.3%
Multi-step tasks where the model has to pick and combine many real tools, connected through MCP, the standard way agents plug into apps and data.
8th of 33. The maker's own number, from Anthropic, via LLM Stats.
- SWE-Bench Pro64.3%
Long changes to real codebases in several languages, often across many files, including code from private and copyleft projects that models are unlikely to have seen.
10th of 50. The maker's own number, from Anthropic, via LLM Stats.
- Humanity's Last Exam36.2%
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.
Independent result from Epoch AI.
- Vending-Bench 2$10,937
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.
8th of 60. Independent result from Epoch AI.
- OSWorld-Verified78%
Ordinary tasks on a real desktop computer using normal apps: a browser, a spreadsheet, an email client, a photo editor, the file manager.
11th of 25. The maker's own number, from Anthropic, via LLM Stats.
- SimpleBench61.7%
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.
35th of 68. Independent result from Epoch AI.
- SimpleQA Verified51.7%
Short factual questions with one checked answer, asked without web search.
29th of 78. Independent result from Epoch AI.
- FrontierMath70.2%
Hundreds of new, unpublished math problems written by mathematicians, from hard undergraduate level up to research, with answers a computer can check.
31st of 76. 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.