Claude Sonnet 4: benchmarks, price and rank
Claude Sonnet 4 from Anthropic is 111th of 123 rated AI models on Spezia's leaderboard. It costs $3 in, $15 out per million tokens.
- 111th of 123 rated models
- Rating 611
- Rated on 5 benchmarks
- Closed
- Released May 22, 2025
- 200,000 tokens of context
Where it is strongest
Not in the top tenth on any benchmark in use.
Where it falls behind
- 101st of 107 on LMArena text
- 87th of 94 on Epoch Capabilities Index
- 61st of 68 on SimpleBench
Price: $3 in, $15 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.
- Humanity's Last Exam7.8%
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.
- SimpleBench45.5%
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.
61st of 68. Independent result from Epoch AI.
- LMArena text1,351
People put a question to two anonymous models and vote for the better answer.
101st of 107. Independent result from LMArena.
- SWE-bench Verified72.7%
Real bug reports and feature requests from popular open-source Python projects. The model has to change the code so the project's hidden tests pass.
26th of 56. The maker's own number, from Anthropic, via LLM Stats.
- GPQA Diamond79.2%
Graduate-level questions in biology, physics and chemistry, written by PhDs so that searching the web does not give the answer away.
88th of 103. Independent result from Epoch AI.
- ARC-AGI-25.9%
Visual puzzles of colored grids, each with a new rule to work out from a few examples.
56th of 69. Independent result from Epoch AI.
- Epoch Capabilities Index141.9
One score built from dozens of benchmarks, set on a common scale so that models tested on different benchmarks can still be compared.
87th of 94. Independent result from Epoch AI.
2 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.