xAI Releases Grok 4.6 with Enhanced Coding and Knowledge Work Capabilities
Summarized by AI from reporting by Matthew Berman, published under our editorial policy.
xAI has launched Grok 4.6, an upgraded model delivering significant performance gains in coding, legal analysis, and general knowledge work. Building on xAI's integration of Cursor data, the model powers both developer platforms and the new consumer-focused Grokbot.
xAI has officially introduced Grok 4.6, a point-upgrade to Grok 4.5 designed to enhance performance across software engineering, reasoning, and broad knowledge tasks. Benchmark results show substantial jumps over its predecessor, including top scores on OpenAI's GDPval benchmark and legal evaluation suite Harvey Lab, alongside a rise to fourth place overall on the Artificial Analysis Intelligence Index. Priced at $2 per million input tokens and $6 per million output tokens, Grok 4.6 offers a competitive balance of speed and efficiency relative to major market competitors.
The development of Grok 4.6 underscores xAI's strategic emphasis on developer-focused workflows and knowledge work following its acquisition of Cursor. According to the company, Grok 4.5 was used to synthesize and curate supervised fine-tuning (SFT) trajectories across software engineering, STEM, and agentic reasoning harnesses to train the new model. In addition to being available via API, OpenRouter, Vercel, and Cloudflare, Grok 4.6 is natively integrated into Cursor and powers Grokbot, a newly launched agent-based consumer tool that abstracts away model selection and technical mechanics to generate documents and artifacts.
xAI's development velocity is set to continue, with company leadership indicating that Grok 4.7 has completed initial training and is undergoing supplemental training using internal SpaceX operational data. As xAI leverages its expansive compute clusters alongside dedicated domain data, the lab continues to solidify its position among the leading developers of frontier artificial intelligence models.