Grok 4.6 Launches with Better Long-Task and Visual Skills

xAI, a US company, launched Grok 4.6 on August 12. It builds on Grok 4.5 to handle longer tasks and complex interactions, like research, analyzing big codebases, and turning ideas into apps.

In tests, Grok 4.6 matches GPT-5.6 Sol on 9 benchmarks. It's now on Cursor and Grok Build, with double credits for the first week. API pricing is $2 per million input tokens and $6 per million output tokens.

The model used extra training with filtered data and lots of reinforcement learning. It's good at turning vague ideas into working prototypes and checks its own work. Safety tests were also the biggest yet.