Claude Fable 5.2 (Max): benchmark results
Provider: Anthropic. Released 2026-11-09. Access: API.
Aggregate Rating 2066 ± 20, rank #1 of 2201 rated models, from 25 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Bug Hunt Bench | 58 | Planted Bugs Fixed (out of 105) | 100 |
| Chess Puzzles (Epoch AI) | 79 | Accuracy (%) | 100 |
| Epoch AI - Ebr Bench | 90.48 | Score | 100 |
| Epoch AI - Furniture Assembly | 90 | Score | 100 |
| Epoch AI - Mystery Game Puzzles | 90 | Score | 100 |
| FrontierMath - Tiers 1-3 (v2) | 95.09 | Accuracy (%, 285 private v2 problems) | 100 |
| FrontierSWE V2 | 69.94 | Mean@5 Score (%) | 100 |
| Humanity's Sixth Sense | 59.5 | Pass@1 (%), all rubric criteria; three attempts/task | 100 |
| SimpleQA Verified | 94 | Accuracy (%) | 100 |
| Bug Hunt Bench - LMS | 35 | Planted Bugs Fixed (out of 60) | 99.6 |
| Bug Hunt Bench - VS Code Extension | 23 | Planted Bugs Fixed (out of 45) | 99.6 |
| FrontierMath - Tier 4 (v2) | 100 | Accuracy (%, 41 private v2 problems) | 99.5 |
Interactive version: theaggregate.ai/model?slug=claude-fable-5-2-max · How It Works · Data refreshed daily, snapshot 2026-11-12.