Claude Haiku 5.5: benchmarks and ranking
Claude Haiku 5.5 (Anthropic) — overall rank: #45 / 153, overall score: 71.6 (90% interval 66.9–75.8), leaderboards: 7. Released: 2026-10-07 · Price (in/out): $0.1 / $0.5
Rank on each leaderboard
| Leaderboard | Rank | Published value |
|---|---|---|
| Artificial Analysis | #21 / 141 | 43.4 |
| LiveBench | #48 / 64 | 72.1 |
| Epoch AI | #20 / 230 | 160.33 |
| ARC Prize (ARC-AGI) | #36 / 95 | 61.3 |
| Scale AI SEAL | Not rated | — |
| Vals AI | #22 / 42 | 58.7 |
| Terminal-Bench | #21 / 53 | 40 |
| EQ-Bench | Not rated | — |
| SimpleBench | #50 / 112 | 59.2 |
| BenchLM.ai | Not rated | — |
Domain scores
- Coding: 78.5
- Agents: 68.9
- Reasoning: 70.4
- Math: 73.7
- Knowledge: 70.1
- Writing: 30.1
- Instructions: 55.1
- Chat & EQ: 30.8
- Business: 78.9
- Japanese: 69.0
- Long context: 58.1