Claude Opus 4.5: benchmarks and ranking
Claude Opus 4.5 (Anthropic) — overall rank: #73 / 153, overall score: 53.4 (90% interval 46.5–60.2), leaderboards: 5. Released: 2025-11-24 · Price (in/out): $5 / $25
Rank on each leaderboard
| Leaderboard | Rank | Published value |
|---|---|---|
| Artificial Analysis | Not rated | — |
| LiveBench | #47 / 64 | 72.6 |
| Epoch AI | #61 / 230 | 151.87 |
| ARC Prize (ARC-AGI) | #50 / 95 | 37.6 |
| Scale AI SEAL | #16 / 44 | 25.2 |
| Vals AI | Not rated | — |
| Terminal-Bench | Not rated | — |
| EQ-Bench | Not rated | — |
| SimpleBench | #41 / 112 | 62 |
| BenchLM.ai | Not rated | — |
Domain scores
- Coding: 59.8
- Agents: 35.3
- Reasoning: 60.6
- Math: 65.4
- Knowledge: 75.2
- Writing: 69.7
- Instructions: 44.9
- Chat & EQ: 41.4
- Business: 81.7
- Japanese: 76.7
- Long context: 54.0