Anthropic
Claude Sonnet 4.6 benchmark results
17 results on 9 benchmarks from 2 sources.
Also listed as: claude-sonnet-4-6 · claude-sonnet-4-6_max · Claude Sonnet 4.6 (max) · claude-sonnet-4-6_high · Claude Sonnet 4.6 (high) · claude-sonnet-4-6_medium · Claude Sonnet 4.6 (medium) · claude-sonnet-4-6_32K · Claude Sonnet 4.6 (32k thinking) · Claude Sonnet 4.6 (no thinking)
All benchmark results
| Benchmarks | Score | Rank | Setting | Source type | Date |
|---|---|---|---|---|---|
| LMArena Text | 1472 | 44 | Overall, style control on; 70971 votes; 95% CI 1468–1476“claude-sonnet-4-6” | Official leaderboard | Oct 8, 2026 |
| LMArena WebDev | 1522 | 48 | Overall; 22866 votes; 95% CI 1517–1528“claude-sonnet-4-6” | Official leaderboard | Oct 8, 2026 |
| LMArena Vision | 1275 | 33 | Overall, style control on; 27429 votes; 95% CI 1269–1281“claude-sonnet-4-6” | Official leaderboard | Oct 2, 2026 |
| GPQA Diamond | 87.4% | 69 | Epoch AI run, accuracy; effort: 32k thinking | Independent run | Feb 20, 2026 |
| GPQA Diamond | 83.3% | 96 | Epoch AI run, accuracy; effort: high | Independent run | Jul 13, 2026 |
| GPQA Diamond | 83.3% | 96 | Epoch AI run, accuracy; effort: medium | Independent run | Jul 13, 2026 |
| GPQA Diamond | 78.8% | 121 | Epoch AI run, accuracy; effort: max | Independent run | Aug 6, 2026 |
| SWE-bench Verified | 75.2% | 14 | Epoch AI run (Epoch scaffold); effort: no thinking | Independent run | Feb 21, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 85.8% | 92 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: 32k thinking | Independent run | Feb 20, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 82.2% | 105 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: medium | Independent run | Jul 13, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 75.6% | 118 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: high | Independent run | Jul 13, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 71.1% | 126 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: max | Independent run | Aug 6, 2026 |
| OSWorld (Verified) | 72.1% | 1 | OSWorld; agent: claude-sonnet-4-6 (100 steps); effort: no thinking | Official leaderboard via Epoch AI | Feb 17, 2026 |
| OSWorld 2.0 | 9.3% | 10 | Binary accuracy; 500 steps; reasoning: medium; effort: medium | Official leaderboard via Epoch AI | Feb 17, 2026 |
| OSWorld 2.0 | 8.3% | 11 | Binary accuracy; 500 steps; reasoning: max; effort: max | Official leaderboard via Epoch AI | Feb 17, 2026 |
| Terminal-Bench 2.0 | 53.4% | 3 | Terminal-Bench 2.0; agent: Simplai Agent; effort: no thinking | Third-party compilation via Epoch AI | May 14, 2026 |
| Terminal-Bench 2.0 | 53.4% | 63 | Terminal-Bench 2.0; agent: Simplai Agent; effort: unknown thinking | Official leaderboard via Epoch AI | Feb 17, 2026 |
FAQ
Where do these Claude Sonnet 4.6 scores come from?
From Epoch AI Benchmarking Hub and LMArena leaderboard dataset. Each row links to the original leaderboard, dataset or report, and vendor-reported numbers are labelled.