Anthropic
Claude Opus 4.6 benchmark results
29 results on 9 benchmarks from 4 sources.
Also listed as: claude-opus-4-6-high · claude-opus-4-6 · claude-opus-4-6_max · Claude Opus 4.6 (max) · claude-opus-4-6_32K · Claude Opus 4.6 (32k thinking) · claude-opus-4-6_64K · Claude Opus 4.6 (64k thinking) · Claude Opus 4.6 (no thinking) · claude-opus-4-6-thinking-max
All benchmark results
| Benchmarks | Score | Rank | Setting | Source type | Date |
|---|---|---|---|---|---|
| LMArena Text | 1504 | 3 | Overall, style control on; 79415 votes; 95% CI 1501–1508“claude-opus-4-6-high” | Official leaderboard | Oct 8, 2026 |
| LMArena Text | 1498 | 7 | Overall, style control on; 84095 votes; 95% CI 1494–1501“claude-opus-4-6” | Official leaderboard | Oct 8, 2026 |
| LMArena WebDev | 1547 | 40 | Overall; 19434 votes; 95% CI 1541–1552“claude-opus-4-6-high” | Official leaderboard | Oct 8, 2026 |
| LMArena WebDev | 1537 | 45 | Overall; 20694 votes; 95% CI 1532–1543“claude-opus-4-6” | Official leaderboard | Oct 8, 2026 |
| LMArena Vision | 1299 | 3 | Overall, style control on; 22328 votes; 95% CI 1293–1306“claude-opus-4-6-high” | Official leaderboard | Oct 2, 2026 |
| LMArena Vision | 1294 | 7 | Overall, style control on; 26869 votes; 95% CI 1288–1301“claude-opus-4-6” | Official leaderboard | Oct 2, 2026 |
| GPQA Diamond | 90.5% | 37 | Epoch AI run, accuracy; effort: 32k thinking | Independent run | Feb 6, 2026 |
| GPQA Diamond | 88.8% | 54 | Epoch AI run, accuracy; effort: 64k thinking | Independent run | Feb 6, 2026 |
| GPQA Diamond | 88.4% | 57 | Epoch AI run, accuracy; effort: max | Independent run | Aug 6, 2026 |
| SWE-bench Verified | 78.7% | 4 | Epoch AI run (Epoch scaffold); effort: no thinking | Independent run | Feb 18, 2026 |
| SWE-bench Verified | 75.6% | 8 | agent: mini-SWE-agent“claude-opus-4-6” | Official leaderboard | Feb 17, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 94.4% | 52 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: 64k thinking | Independent run | Feb 6, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 93.1% | 62 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: 32k thinking | Independent run | Feb 6, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 91.1% | 67 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: max | Independent run | Aug 6, 2026 |
| Humanity's Last Exam | 34.4% | 11 | Compiled by Epoch AI; effort: max | Third-party compilation via Epoch AI | Feb 5, 2026 |
| Humanity's Last Exam | 19.0% | 24 | Compiled by Epoch AI; effort: no thinking | Third-party compilation via Epoch AI | Feb 5, 2026 |
| τ²-bench | 27.3% | 15 | banking_knowledge domain, pass^1; effort: max; submitted by Sierra; v1.0.1 | Official leaderboard | May 6, 2026 |
| Terminal-Bench 2.0 | 79.8% | 8 | Terminal-Bench 2.0; agent: ForgeCode; effort: unknown thinking | Official leaderboard via Epoch AI | Mar 12, 2026 |
| Terminal-Bench 2.0 | 76.4% | 13 | Terminal-Bench 2.0; agent: Meta-Harness; effort: unknown thinking | Official leaderboard via Epoch AI | Feb 5, 2026 |
| Terminal-Bench 2.0 | 75.3% | 16 | Terminal-Bench 2.0; agent: Capy; effort: unknown thinking | Official leaderboard via Epoch AI | Feb 5, 2026 |
| Terminal-Bench 2.0 | 74.7% | 19 | Terminal-Bench 2.0; agent: Terminus-KIRA; effort: unknown thinking | Official leaderboard via Epoch AI | Feb 5, 2026 |
| Terminal-Bench 2.0 | 72.1% | 21 | Terminal-Bench 2.0; agent: MAYA-V2; effort: unknown thinking | Official leaderboard via Epoch AI | Feb 5, 2026 |
| Terminal-Bench 2.0 | 71.9% | 22 | Terminal-Bench 2.0; agent: TongAgents; effort: unknown thinking | Official leaderboard via Epoch AI | Feb 5, 2026 |
| Terminal-Bench 2.0 | 69.9% | 24 | Terminal-Bench 2.0; agent: Droid; effort: no thinking | Official leaderboard via Epoch AI | Feb 5, 2026 |
| Terminal-Bench 2.0 | 69.9% | 25 | Terminal-Bench 2.0; agent: Droid; effort: unknown thinking | Official leaderboard via Epoch AI | Feb 5, 2026 |
| Terminal-Bench 2.0 | 66.9% | 28 | Terminal-Bench 2.0; agent: Crux; effort: unknown thinking | Official leaderboard via Epoch AI | Feb 5, 2026 |
| Terminal-Bench 2.0 | 66.5% | 29 | Terminal-Bench 2.0; agent: Mux; effort: unknown thinking | Official leaderboard via Epoch AI | Feb 5, 2026 |
| Terminal-Bench 2.0 | 62.9% | 39 | Terminal-Bench 2.0; agent: Terminus 2; effort: unknown thinking | Official leaderboard via Epoch AI | Feb 5, 2026 |
| Terminal-Bench 2.0 | 58.0% | 52 | Terminal-Bench 2.0; agent: Claude Code; effort: unknown thinking | Official leaderboard via Epoch AI | Feb 5, 2026 |
FAQ
Where do these Claude Opus 4.6 scores come from?
From Epoch AI Benchmarking Hub, LMArena leaderboard dataset, 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.