Claude Fable 5.1 (High): benchmark results

Provider: Anthropic. Released 2026-09-01. Access: API.

Unified ELO 1795 ± 1, rank #8 of 1761 rated models, from 12 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Computer Anthology Terminal Tasks (Claude Code)55.2pass@1 (%)100
Computer Anthology Terminal Tasks (Terminus-2)62.4pass@1 (%)100
WeirdML92.3Average Score99.4
Creative Writing (Lechmazur)4.1Mean Score98.9
Epoch AI - Critpt30.29Score98.2
Agents on Rails92.1Successful Runs (%)96.9
Epoch AI - Apex Agents44.4Score96.8
Vals AI Terminal-Bench 2.185.02Accuracy (%)96.7
ARC-AGI-288.75Accuracy (%)95.5
Deep20Bench12.11Questions to Solve (mean, lower is better)93.8
ARC-AGI-196Accuracy (%)92
CursorBench 3.169.4Score (%)91.8

Interactive version: theaggregate.ai/model?slug=claude-fable-5-1-high · How It Works · Data refreshed daily, snapshot 2026-09-05.