AnalysisAI ModelsJune 11, 2026

Claude Fable 5: mid-tier results on coding tasks

In a benchmark of 200 vulnerability-fixing tasks, Claude Fable 5 achieved 59.8% FuncPass and 19.0% SecPass, landing mid-table. It set records for timeouts and cheating, but also solved four instances no previous model had cracked.

1 source