Claude Opus 4.6: benchmarks and ranking
Claude Opus 4.6 (Anthropic) — overall rank: #50 / 153, overall score: 67.7 (90% interval 61.6–73.3), leaderboards: 6. Released: 2025-12-17 · Price (in/out): $5 / $25
Rank on each leaderboard
| Leaderboard | Rank | Published value |
|---|---|---|
| Artificial Analysis | Not rated | — |
| LiveBench | #39 / 64 | 74.5 |
| Epoch AI | #39 / 230 | 156.33 |
| ARC Prize (ARC-AGI) | #30 / 95 | 69.2 |
| Scale AI SEAL | #11 / 44 | 34.44 |
| Vals AI | Not rated | — |
| Terminal-Bench | Not rated | — |
| EQ-Bench | #17 / 32 | 1210 |
| SimpleBench | #32 / 112 | 67.6 |
| BenchLM.ai | Not rated | — |
Domain scores
- Coding: 67.2
- Agents: 59.9
- Reasoning: 77.3
- Math: 67.6
- Knowledge: 82.1
- Writing: 75.1
- Instructions: 45.0
- Chat & EQ: 52.1
- Business: 76.8
- Japanese: 84.6
- Vision: 74.1
- Long context: 88.7