AnthropicClaude Sonnet
Claude 3.7 Sonnet
Anthropic · Claude Sonnet. Released Feb 24, 2025.
Claude 3.7 Sonnet is a model from Anthropic in the Claude Sonnet family, released Feb 24, 2025. evals.report tracks 33 reported Claude 3.7 Sonnet benchmark scores across SWE-bench Verified, GPQA Diamond, FrontierMath, AIME (OTIS Mock), LMArena, Epoch Capabilities Index, Aider Polyglot, MMLU-Pro, and 25 more — each shown with its benchmark, metric, source status, and date, and never combined into a single ranking.
33 results
Benchmark results 33
Compare this model| Benchmark | Category | Score | Metric | Status | Date | |
|---|---|---|---|---|---|---|
| SWE-bench Verified | Coding | 61.0% | % resolved | Official | Feb 24, 2025 | Details |
| GPQA Diamond | Reasoning | 78.5% | accuracy | Official | Feb 24, 2025 | Details |
| FrontierMath | Reasoning | 4.14% | accuracy | Official | Feb 24, 2025 | Details |
| AIME (OTIS Mock) | Reasoning | 57.8% | accuracy | Official | Feb 24, 2025 | Details |
| LMArena | Chat preference | 1299 | source-defined rating | Official | Feb 24, 2025 | Details |
| Epoch Capabilities Index | Reasoning | 142.0 | Index | Official | Feb 24, 2025 | Details |
| Aider Polyglot | Coding | 64.9% | % correct | Official | Feb 24, 2025 | Details |
| MMLU-Pro | Reasoning | 83.7% | accuracy | Verified | Feb 24, 2025 | Details |
| GAIA: A Benchmark for General AI Assistants | Agents | 43.9% | accuracy | Unverified | Feb 24, 2025 | Details |
| τ²-bench (Telecom) | Tool use | 49% | pass^1 | Verified | Feb 24, 2025 | Details |
| GDPval | Agents | 1048 | Elo | Official | Feb 24, 2025 | Details |
| LiveCodeBench | Coding | 39.4% | Pass@1 | Unverified | Feb 24, 2025 | Details |
| METR Task-Completion Time Horizons | Agents | 60.4 min | 50% time horizon | Official | Feb 24, 2025 | Details |
| SciCode | Coding | 40.3% | accuracy | Unverified | Feb 24, 2025 | Details |
| MMMU (Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark) | Multimodal | 75.0% | accuracy | Unverified | Feb 24, 2025 | Details |
| MultiChallenge | Reasoning | 51.58% | accuracy | Verified | Feb 24, 2025 | Details |
| Arena-Hard-Auto v2.0 | Chat preference | 59.8% | % win rate | Official | Feb 24, 2025 | Details |
| EQ-Bench Creative Writing v3 | Chat preference | 1395 | Elo | Verified | Feb 24, 2025 | Details |
| Design Arena | Chat preference | 1231 | Elo | Verified | Feb 24, 2025 | Details |
| MASK (Model Alignment between Statements and Knowledge) | Other | 82.13 | Honesty score | Verified | Feb 24, 2025 | Details |
| MCP-Universe | Tool use | 24.24% | Overall Success Rate | Verified | Feb 24, 2025 | Details |
| BigCodeBench | Coding | 33.8% | calibrated Pass@1 | Verified | Feb 24, 2025 | Details |
| SWE-bench Multilingual | Coding | 43% | % resolved | Verified | Feb 24, 2025 | Details |
| SWE-bench Multimodal | Coding | 31.33% | % resolved | Verified | Feb 24, 2025 | Details |
| EnigmaEval | Reasoning | 4.23% | accuracy | Verified | Feb 24, 2025 | Details |
| ZeroBench | Multimodal | 1.0% (pass@1) | accuracy | Verified | Feb 24, 2025 | Details |
| PutnamBench | Reasoning | 0 | Problems solved | Verified | Feb 24, 2025 | Details |
| Gray Swan Arena (Agent Red-Teaming / Indirect Prompt Injection) | Agents | 1.61% | Attack Success Rate (ASR) | Verified | Feb 24, 2025 | Details |
| PolyMath: Evaluating Mathematical Reasoning in Multilingual Contexts | Reasoning | 33.5 | Difficulty-Weighted Accuracy (DW-ACC) | Verified | Feb 24, 2025 | Details |
| Online-Mind2Web | Agents | 39.33% | Task success rate | Verified | Feb 24, 2025 | Details |
| WebArena | Agents | 52.0% | Task success rate | Verified | Feb 24, 2025 | Details |
| GSO: Software Optimization Benchmark for SWE-Agents | Coding | 3.8% | Opt@1 | Official | Feb 24, 2025 | Details |
| MultiNRC | Reasoning | 27.77% | accuracy | Official | Feb 24, 2025 | Details |