AnalysisAI ModelsJuly 16, 2026

Fable 5 and GPT-5.6 Sol solve Baba Is You levels in new benchmark

A new open-source benchmark, Baba is Harbor, ports Baba Is You levels to the Harbor agentic RL framework. Fable 5 and GPT-5.6 Sol solved Stage 0 and Stage 1, but the post questions the cost.

1 source