Fable 5.1 tops Bug Hunt Bench, beating GPT-5.6 with better speed and cost

PawelHuryn · x · 2026-09-02

Fable 5.1 achieved the top score on the Bug Hunt Bench, fixing 43 out of 105 planted bugs across 2 real repositories, outperforming GPT-5.6 Sol (42/105) and Grok 4.6 (27/105). It is 2.24x faster than GPT-5.6 Sol and 26% cheaper than the previous Fable 5 version. The poster noted that while Fable 5 felt good to use, it wasn't the best coder, making the performance of Fable 5.1 surprising.

Related event: Anthropic's Fable 5.1 Tops Bug Hunt Bench, Beating GPT-5.6(4 posts)→

Original post →

More from Models

Models channel →