xAI has launched Grok 4.6, an upgraded AI model focused on long-running agents and ambitious interactive and visual work. Building on Grok 4.5, it sustains complex multi-step tasks such as research, codebase analysis, and turning ideas into polished applications.
Grok 4.6 matches frontier performance on key benchmarks, including the Artificial Analysis Intelligence Index (tied with GPT-5.6 Sol). It shows strong gains in agentic coding and knowledge-work evaluations.
The model is available now in Cursor and Grok Build, with 2× included usage for the first week. It is also accessible via the xAI API and partners including OpenRouter, Vercel, and Cloudflare. Pricing starts at $2 per million input tokens and $6 per million output tokens.
Training improvements include a longer supplemental run, curated high-quality data, and extensive agentic reinforcement learning across coding, STEM, and specialized domains. Safety systems have been updated to match the model’s expanded capabilities.
Grok 4.6 is particularly strong at converting broad product ideas into working first versions with better structure, visual design, and self-verification on longer trajectories.
Source: x.ai news