xAI ships Grok 4.6 to tackle long-running agent tasks

xAI's latest model matches top frontier scores while staying focused on long, multi-step coding and design tasks.

xAI released Grok 4.6 today. According to the announcement, the updated model targets long-running projects—like building full applications, editing large codebases, and running research tasks over many steps. It scored 61 on the Artificial Analysis Intelligence Index, matching OpenAI's GPT-5.6 Sol Max and topping the previous Grok 4.5 score of 56.

Why it matters: Complex coding projects usually stall when models forget context after ten prompts. xAI trained Grok 4.6 to self-test and verify its own code along the way. That extra checking produces cleaner first passes on full software builds and visual layouts instead of broken prototypes.

Try it: You can use it today in Cursor, Grok Build, OpenRouter, Vercel, Cloudflare, and the xAI API. Base pricing runs $2 per million input tokens and $6 per million output tokens. Cursor and Grok Build are offering double included usage for the first week.

Keep an eye on your API bill if speed is your priority—the fast variant costs twice as much per token.