xAI Releases Grok 4.6: Focuses on Long-Running Agents, Matches GPT-5.6

scaling01 · x · 2026-08-12

xAI has officially announced the release of Grok 4.6, with a particular focus on improving long-running agents and complex interactive visual work. The model is designed to maintain stability across multi-step tasks like research analysis and codebase maintenance.

According to the announcement, Grok 4.6 achieves frontier intelligence in agentic coding and knowledge work benchmarks, matching GPT-5.6 Sol on the Artificial Analysis Intelligence Index. It is now available in Cursor and Grok Build with 2x included usage for the first week. The training involved a longer supplemental training run than Grok 4.5 and used Grok 4.5 to regenerate SFT trajectories.

Related event: xAI Officially Releases Grok 4.6 with API Access(9 posts)→

Original post →

More from coding & agent

coding & agent channel →