NEWS: Grok 4.6 beats GPT-5.6 on Devin's coding benchmark.
Cognition, the company behind the AI engineer Devin, added Grok 4.6 and shared an independent test. On its FrontierCode 1.1 benchmark, Grok 4.6 scored 61.3, ahead of GPT-5.6 Sol at 60.6 and a clear jump over Grok 4.5 at 56.5.
Cognition said Grok 4.6 "marks a significant improvement over Grok 4.5, surpassing GPT-5.6 Sol, behind only Opus 5 and Fable 5." Those two Claude models led the chart at 63.6 and 64.9.
Grok 4.6 is now available in Devin.
Source: Cognition on X, August 2026