xAI Releases Grok 4.6 with Enhanced Agentic and Visual Capabilities
Original: Grok 4.6
Why This Matters
Grok 4.6 signals intensifying competition at the frontier of agentic AI coding and knowledge work models.
xAI released Grok 4.6 on August 12, 2026, building on Grok 4.5 with a focus on long-running agents and interactive/visual work. The model matches GPT-5.6 Sol on the Artificial Analysis Intelligence Index with a score of 61, and is available in Cursor and Grok Build with 2x included usage for the first week.
xAI released Grok 4.6 on August 12, 2026, positioning it as an upgrade to Grok 4.5 with a focus on long-running agentic tasks and interactive/visual work. The model achieves a score of 61 on the Artificial Analysis Intelligence Index — a composite of nine benchmarks — matching GPT-5.6 Sol and trailing only Fable 5 Max (62). It outperforms Grok 4.5 (56) across benchmarks including GDPVal-AA, DeepSWE 1.1, CursorBench 3.2, and FrontierCode 1.1.
Training involved a longer supplemental run than Grok 4.5, using curated model-generated data for reasoning and advanced technical concepts, high-quality engineering data, and an improved optimizer. xAI used Grok 4.5 to regenerate SFT trajectories across reasoning, agent harnesses, and domains such as STEM, software engineering, and knowledge work. Agentic RL training covered knowledge work, general coding, kernel optimization, web development, and computer-aided design.
xAI highlights the model's ability to turn broad product ideas into working first versions, with improved self-testing behavior on longer trajectories. Safety evaluations were the company's widest-ever pre-deployment suite. Grok 4.6 is available now in Cursor and Grok Build, with 2x included usage offered during the first week.