Claude Haiku 5.5 (Max): benchmark results
Provider: Anthropic. Access: API.
Unified ELO 1742 ± 28, rank #246 of 2096 rated models, from 28 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| LiveBench Math Comp | 98.04 | Score | 98.5 |
| LiveBench Theory of Mind | 86.54 | Score | 92.6 |
| CursorBench 4.0 | 48.4 | Score (%) | 86.8 |
| LiveBench Python | 60 | Score | 72.1 |
| LiveBench JavaScript | 68.18 | Score | 69.1 |
| LiveBench TypeScript | 43.33 | Score | 55.9 |
| Bug Hunt Bench - VS Code Extension | 10 | Planted Bugs Fixed (out of 45) | 54.9 |
| Bug Hunt Bench | 21.5 | Planted Bugs Fixed (out of 105) | 51.5 |
| Bug Hunt Bench - LMS | 11.5 | Planted Bugs Fixed (out of 60) | 49.5 |
| LiveBench Code Completion | 78.26 | Score | 49.3 |
| LiveBench Code Generation | 77.47 | Score | 47.8 |
| LiveBench Zebra Puzzle | 94 | Score | 41.9 |
Interactive version: theaggregate.ai/model?slug=claude-haiku-5-5-max · How It Works · Data refreshed daily, snapshot 2026-10-08.