xAI's Grok 4.6 Catches Up with Top AI Models and Underbids Them Significantly on Price

According to the Artificial Analysis Intelligence Index, xAI's new model Grok 4.6 scores 61 points, matching OpenAI's GPT-5.6 Sol and behind only Anthropic's Claude Opus 5 (63) and Claude Fable 5 (62). Compared to its predecessor Grok 4.5, the model improves by five points. Grok 4.6 excels particularly at agentic tasks, which involve autonomously executed multi-step workflows. On the GDPval-AA v2 benchmark, which measures real computer-based knowledge work, it ranks second with an Elo score of 1753, behind Claude Opus 5. It completes complex tasks in about 53 steps, while Claude Opus 5 needs about 103. The price remains at 2/6 dollars per million tokens, more than 60 percent below Claude Opus 5 (5/25 dollars) and GPT-5.6 Sol (5/30 dollars). Grok 4.6 is immediately available via the API, in Cursor, Grok Build, and through partners such as OpenRouter, Vercel, and Cloudflare. In the first week, x.ai offers double usage in Grok Build and Cursor.