Claude Opus 5.5: benchmarks and ranking
Claude Opus 5.5 (Anthropic) — overall rank: #4 / 153, overall score: 90.1 (90% interval 88.2–91.8), leaderboards: 9. Released: 2026-09-22 · Price (in/out): $4 / $20
Rank on each leaderboard
| Leaderboard | Rank | Published value |
|---|---|---|
| Artificial Analysis | #1 / 141 | 57.6 |
| LiveBench | #2 / 64 | 83.2 |
| Epoch AI | #5 / 230 | 168.67 |
| ARC Prize (ARC-AGI) | #6 / 95 | 93.3 |
| Scale AI SEAL | Not rated | — |
| Vals AI | #5 / 42 | 70.1 |
| Terminal-Bench | #2 / 53 | 65.76 |
| EQ-Bench | #2 / 32 | 1339.5 |
| SimpleBench | #6 / 112 | 88.4 |
| BenchLM.ai | #6 / 24 | 82.8 |
Domain scores
- Coding: 93.3
- Agents: 93.3
- Reasoning: 88.6
- Math: 84.5
- Knowledge: 88.3
- Writing: 82.9
- Instructions: 69.3
- Chat & EQ: 89.6
- Business: 88.7
- Japanese: 87.8
- Vision: 85.8
- Long context: 63.1