Grok 4.6 challenges OpenAI, Anthropic's top AI models
What's the story
SpaceXAI has launched its latest artificial intelligence (AI) model, Grok 4.6. It is specifically designed for advanced coding and knowledge work. The Elon Musk-led company claims that Grok 4.6 can now match OpenAI's GPT-5.6 Sol in reasoning and stand toe-to-toe with Anthropic's Claude Fable 5. This comes after SpaceXAI's recent acquisition of Cursor and the launch of Grok Bot, signaling its intent to take on top players in the industry.
Enhanced features
Grok 4.6 matches GPT-5.6 Sol's performance
Grok 4.6 has been trained on a wealth of carefully curated data, making it adept at handling complex multistep tasks, understanding new codebases, and self-verifying its work.
The model also received extra training for software engineering and kernel optimization.
It scored an impressive 61 on the Artificial Analysis Intelligence Index, matching GPT-5.6 Sol's performance while outperforming its predecessor in key benchmarks.
Access details
Grok 4.6 now available on multiple platforms
Grok 4.6 is now available on platforms like Cursor, Grok Build, OpenRouter, and Vercel. The pricing starts at $2 per million input tokens and $6 per million output tokens. A faster version costs twice as much.
The model also comes with enhanced security features for tasks such as vulnerability patching and design work.
This launch marks a major milestone for SpaceXAI in its quest to establish itself as a key player in the AI space.