Claude Fable 5 (Max) — benchmark results

Claude Fable 5 evaluated at the max reasoning-effort setting. Provider: Anthropic. Released 2026-06-09. Access: API.

Unified ELO 2173 ± 37, rank #6 of 1776 rated models, from 29 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
CursorBench 3.170.5Score (%)100
Epoch AI - Cursorbench72.9Score100
Epoch AI - Scicode60.19Score100
FrontierMath - Tier 4 (v2)87.8Accuracy (%, 41 private v2 problems)100
MathArena - ARXIV April70.73Accuracy (%)100
MathArena Arxiv87.5Accuracy (%)100
MedScribe88.52Score (self-reported)100
Vals Index75.14Score (self-reported)100
Vals Multimodal Index74.15Score (self-reported)100
WeirdML91.94Average Score100
CritPt28.6Accuracy (self-reported)99.2
ZeroBench23Score (%)99.2

Interactive version: theaggregate.ai/model?slug=claude-fable-5-max · How the rankings work · Data refreshed daily, snapshot 2026-07-22.