SpaceXAI's latest model, Grok 4.6, has achieved a score of 61 on the Artificial Analysis Intelligence Index, matching OpenAI's GPT-5.6 Sol. This places it in the top tier of models, with only Anthropic's Claude Opus 5 (63) and Claude Fable 5 (62) scoring higher. The AA Index combines multiple benchmarks into a single score, highlighting Grok 4.6's strong performance across various metrics. The model represents a five-point improvement over its predecessor, Grok 4.5, demonstrating significant progress in its capabilities.
Grok 4.6 excels in agentic tasks, where models independently execute multi-step workflows. On the GDPval-AA v2 benchmark, which assesses real-world knowledge work, it ranks second with an Elo score of 1,753. This places it just behind Claude Opus 5, which requires about 103 steps to complete complex tasks, compared to Grok 4.6's 53 steps. The model's pricing remains at $2 per million tokens for standard use and $6 for premium access, making it significantly more affordable than competitors like Claude Opus 5 ($5/$25) and GPT-5.6 Sol ($5/$30).
According to the source, Grok 4.6 is now available through the API, Cursor, Grok Build, and partners such as OpenRouter, Vercel, and Cloudflare. For the first week, x.ai is offering double the usage quota in Grok Build and Cursor. Source: thedecoder
