Claude Fable 5.2 (xHigh): benchmark results
Provider: Anthropic. Released 2026-11-09. Access: API.
Aggregate Rating 2033 ± 36, rank #2 of 2201 rated models, from 15 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| ARC-AGI-3 | 81.17 | Accuracy (%) | 100 |
| Conceptual Reasoning Index | 79.89 | Chance-Corrected Score (0-100) | 100 |
| Conceptual Reasoning Index - Argument Evaluation (LMCA) | 70.61 | Chance-Corrected Score (0-100) | 100 |
| Conceptual Reasoning Index - Consistency (ACCoRD) | 89.39 | Chance-Corrected Score (0-100) | 100 |
| DuelLab GameBench 2 - Reasoning Variants | 99.11 | GameBench2 normalized score (0–100) | 100 |
| Epoch AI - Critpt | 35.71 | Score | 99.6 |
| Conceptual Reasoning Index - Decision Theory (DTBench) | 98.22 | Chance-Corrected Score (0-100) | 99.4 |
| Epoch AI - Dtbench | 98.93 | Score | 99.4 |
| RuneBench | 6.98 | Mean ln(1 + peak XP/min) across sixteen 30-minute skills; hi | 99 |
| ARC-AGI-2 | 93.75 | Accuracy (%) | 98.7 |
| ARC-AGI-1 | 98.5 | Accuracy (%) | 98.1 |
| Bug Hunt Bench | 49 | Planted Bugs Fixed (out of 105) | 96.9 |
Interactive version: theaggregate.ai/model?slug=claude-fable-5-2-xhigh · How It Works · Data refreshed daily, snapshot 2026-11-12.