Anthropic
Claude Opus 5 benchmark results
15 results on 7 benchmarks from 3 sources.
Also listed as: claude-opus-5-high · claude-opus-5 · claude-opus-5_low · Claude Opus 5 (low) · claude-opus-5_max · Claude Opus 5 (max) · claude-opus-5_xhigh · Claude Opus 5 (xhigh) · claude-opus-5_high · Claude Opus 5 (high)
All benchmark results
| Benchmarks | Score | Rank | Setting | Source type | Date |
|---|---|---|---|---|---|
| LMArena Text | 1490 | 13 | Overall, style control on; 62934 votes; 95% CI 1486–1494“claude-opus-5-high” | Official leaderboard | Oct 8, 2026 |
| LMArena WebDev | 1657 | 12 | Overall; 22340 votes; 95% CI 1651–1662“claude-opus-5-high” | Official leaderboard | Oct 8, 2026 |
| LMArena Vision | 1289 | 14 | Overall, style control on; 15900 votes; 95% CI 1282–1296“claude-opus-5-high” | Official leaderboard | Oct 2, 2026 |
| GPQA Diamond | 93.9% | 13 | Epoch AI run, accuracy; effort: max | Independent run | Jul 24, 2026 |
| GPQA Diamond | 92.9% | 21 | Epoch AI run, accuracy | Independent run | Aug 6, 2026 |
| GPQA Diamond | 87.9% | 62 | Epoch AI run, accuracy; effort: low | Independent run | Aug 6, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 98.9% | 20 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: max | Independent run | Jul 24, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 97.8% | 31 | Epoch AI run, OTIS Mock AIME 2024–2025 | Independent run | Aug 6, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 93.3% | 56 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: low | Independent run | Aug 6, 2026 |
| OSWorld 2.0 | 31.4% | 1 | Binary accuracy; 500 steps; reasoning: max; effort: max | Official leaderboard via Epoch AI | Jul 24, 2026 |
| OSWorld 2.0 | 30.2% | 2 | Binary accuracy; 500 steps; reasoning: xhigh; effort: xhigh | Official leaderboard via Epoch AI | Jul 24, 2026 |
| OSWorld 2.0 | 29.0% | 3 | Binary accuracy; 500 steps; reasoning: high; effort: high | Official leaderboard via Epoch AI | Jul 24, 2026 |
| OSWorld 2.0 | 25.3% | 5 | Binary accuracy; 500 steps; reasoning: medium; effort: medium | Official leaderboard via Epoch AI | Jul 24, 2026 |
| OSWorld 2.0 | 22.3% | 6 | Binary accuracy; 500 steps; reasoning: low; effort: low | Official leaderboard via Epoch AI | Jul 24, 2026 |
| τ²-bench | 48.7% | 2 | banking_knowledge domain, pass^1; effort: max; submitted by Sierra; v1.0.1 | Official leaderboard | Aug 3, 2026 |
FAQ
Where do these Claude Opus 5 scores come from?
From Epoch AI Benchmarking Hub, LMArena leaderboard dataset, and τ²-bench leaderboard submissions. Each row links to the original leaderboard, dataset or report, and vendor-reported numbers are labelled.