Anthropic
Claude 3.7 Sonnet benchmark results
28 results on 10 benchmarks from 5 sources.
Closed Released: Feb 24, 2025 Updated Oct 22, 2026
Also listed as: claude-3-7-sonnet-20250219-thinking-32k · claude-3-7-sonnet-20250219 · claude-3-7-sonnet-20250219_64K · Claude 3.7 Sonnet (64k thinking) · claude-3-7-sonnet-20250219_32K · Claude 3.7 Sonnet (32k thinking) · claude-3-7-sonnet-20250219_16K · Claude 3.7 Sonnet (16k thinking) · claude-3-7-sonnet-20250219_unknown · Claude 3.7 Sonnet (unknown thinking)
All benchmark results
| Benchmarks | Score | Rank | Setting | Source type | Date |
|---|---|---|---|---|---|
| LMArena Text | 1388 | 177 | Overall, style control on; 38162 votes; 95% CI 1384–1392“claude-3-7-sonnet-20250219-thinking-32k” | Official leaderboard | Oct 8, 2026 |
| LMArena Text | 1372 | 195 | Overall, style control on; 42573 votes; 95% CI 1368–1376“claude-3-7-sonnet-20250219” | Official leaderboard | Oct 8, 2026 |
| LMArena Vision | 1193 | 89 | Overall, style control on; 1438 votes; 95% CI 1177–1208“claude-3-7-sonnet-20250219-thinking-32k” | Official leaderboard | Oct 2, 2026 |
| LMArena Vision | 1175 | 99 | Overall, style control on; 4519 votes; 95% CI 1166–1185“claude-3-7-sonnet-20250219” | Official leaderboard | Oct 2, 2026 |
| GPQA Diamond | 79.7% | 117 | Epoch AI run, accuracy; effort: 64k thinking | Independent run | May 26, 2025 |
| GPQA Diamond | 76.8% | 129 | Epoch AI run, accuracy; effort: 32k thinking | Independent run | Mar 10, 2025 |
| GPQA Diamond | 76.8% | 129 | Epoch AI run, accuracy; effort: 16k thinking | Independent run | Feb 26, 2025 |
| GPQA Diamond | 66.0% | 172 | Epoch AI run, accuracy | Independent run | Feb 24, 2025 |
| MMLU-Pro | 84.0% | 40 | Overall accuracy; data source: Self-Reported“Claude-3.7-Sonnet-Thinking” | Vendor-reported | Mar 11, 2026 |
| SWE-bench Verified | 61.0% | 29 | Epoch AI run (Epoch scaffold) | Independent run | Feb 4, 2026 |
| SWE-bench Verified | 66.4% | 34 | agent: Aime-coder v1“claude-3-7-sonnet-20250219” | Official leaderboard | May 14, 2025 |
| SWE-bench Verified | 62.4% | 46 | agent: SWE-agent“claude-3-7-sonnet-20250219” | Official leaderboard | Feb 25, 2025 |
| SWE-bench Verified | 52.8% | 65 | agent: mini-SWE-agent“claude-3-7-sonnet-20250219” | Official leaderboard | Jul 20, 2025 |
| AIME (OTIS Mock AIME 2024–2025) | 57.8% | 152 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: 64k thinking | Independent run | Mar 13, 2025 |
| AIME (OTIS Mock AIME 2024–2025) | 53.3% | 161 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: 32k thinking | Independent run | Mar 12, 2025 |
| AIME (OTIS Mock AIME 2024–2025) | 46.7% | 172 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: 16k thinking | Independent run | Feb 26, 2025 |
| AIME (OTIS Mock AIME 2024–2025) | 21.9% | 205 | Epoch AI run, OTIS Mock AIME 2024–2025 | Independent run | Feb 25, 2025 |
| MATH Level 5 | 91.2% | 20 | Epoch AI run, accuracy; effort: 64k thinking | Independent run | Mar 13, 2025 |
| MATH Level 5 | 90.0% | 22 | Epoch AI run, accuracy; effort: 32k thinking | Independent run | Mar 12, 2025 |
| MATH Level 5 | 86.3% | 28 | Epoch AI run, accuracy; effort: 16k thinking | Independent run | Feb 26, 2025 |
| MATH Level 5 | 68.2% | 44 | Epoch AI run, accuracy | Independent run | Feb 24, 2025 |
| Humanity's Last Exam | 8.0% | 38 | Compiled by Epoch AI; effort: unknown thinking | Third-party compilation via Epoch AI | Feb 24, 2025 |
| OSWorld (Verified) | 35.8% | 8 | OSWorld; agent: claude-3-7-sonnet-20250219 (50 steps) | Official leaderboard via Epoch AI | Jul 27, 2025 |
| OSWorld (Verified) | 35.6% | 9 | OSWorld; agent: claude-3-7-sonnet-20250219 (100 steps) | Official leaderboard via Epoch AI | Jul 27, 2025 |
| OSWorld (Verified) | 27.1% | 13 | OSWorld; agent: claude-3-7-sonnet-20250219 (15 steps) | Official leaderboard via Epoch AI | Jul 27, 2025 |
| τ²-bench | 64.2% | 12 | airline domain, pass^1; submitted by Sierra; v0.1.3“Claude-3.7-Sonnet” | Official leaderboard | Jun 9, 2025 |
| τ²-bench | 72.1% | 23 | retail domain, pass^1; submitted by Sierra; v0.1.3“Claude-3.7-Sonnet” | Official leaderboard | Jun 9, 2025 |
| τ²-bench | 49.0% | 19 | telecom domain, pass^1; submitted by Sierra; v0.1.3“Claude-3.7-Sonnet” | Official leaderboard | Jun 9, 2025 |
FAQ
Where do these Claude 3.7 Sonnet scores come from?
From Epoch AI Benchmarking Hub, LMArena leaderboard dataset, MMLU-Pro leaderboard (TIGER-Lab), SWE-bench leaderboards (swebench.com), and τ²-bench leaderboard submissions. Each row links to the original leaderboard, dataset or report, and vendor-reported numbers are labelled.