Claude Sonnet 5: benchmarks, price and rank
Claude Sonnet 5 from Anthropic is 41st of 123 rated AI models on Spezia's leaderboard, strongest on Humanity's Last Exam and SWE-bench Verified. It costs $2 in, $10 out per million tokens.
- 41st of 123 rated models
- Rating 1,413
- Rated on 17 benchmarks
- Closed
- Released Jun 30, 2026
- 1 million tokens of context
Where it is strongest
- 8th of 74 on Humanity's Last Exam
- 3rd of 56 on SWE-bench Verified
Where it falls behind
- 20th of 20 on CursorBench
- 11th of 13 on OfficeQA Pro
- 58th of 78 on SimpleQA Verified
Price: $2 in, $10 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-Agents54.5%
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.
27th of 56. Independent result from Epoch AI.
- Terminal-Bench 4.012.4%
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.
- BrowseComp84.7%
Questions whose answers are hidden deep on the web and take many searches and pages to track down.
12th of 56. The maker's own number, from Anthropic, via LLM Stats.
- SWE-Bench Pro63.2%
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.
12th of 50. The maker's own number, from Anthropic, via LLM Stats.
- Humanity's Last Exam57.4%
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.
8th of 74. The maker's own number, from Anthropic, via LLM Stats.
- Vending-Bench 2$6,378
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.
22nd of 60. Independent result from Epoch AI.
- OSWorld-Verified81.2%
Ordinary tasks on a real desktop computer using normal apps: a browser, a spreadsheet, an email client, a photo editor, the file manager.
6th of 25. The maker's own number, from Anthropic, via LLM Stats.
- Toolathlon54.3%
Long tasks across many real apps and their tools, such as email, calendars, spreadsheets, code repositories and online stores.
14th of 37. The maker's own number, from Anthropic, via LLM Stats.
- CursorBench34.1%
Ambiguous coding tasks that span many files, taken from real sessions in the Cursor code editor.
20th of 20. Independent result from Epoch AI.
- DeepSWE 1.154%
Software engineering tasks run in the same simple agent setup for every model.
Independent result from LLM Stats.
14 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.