FuriaHubSpaceXAI releases Grok 4.6 for long-running agents
The August 12 launch post described a post-training upgrade on Grok 4.5, a 500K context window, a new xhigh reasoning tier, and same-day availability in Cursor and Grok Build.
On August 12 SpaceXAI — the company formerly called xAI — released Grok 4.6. The launch post said the model builds on Grok 4.5 with a focus on long-running agents and more ambitious interactive and visual work, and that it is available the same day in Cursor and Grok Build, with 2× included usage in those products for the first week.
Published specs
Developer docs list the API name grok-4-6, a 500,000-token context window, a February 1, 2026 knowledge cutoff, text and image input with text-only output, and reasoning_effort values of low, medium, high (default), and a new xhigh. MarkTechPost, citing SpaceXAI, said this is a post-training upgrade on the 4.5 foundation rather than a larger base model, and that the model scored 61 on the Artificial Analysis Intelligence Index, tied with GPT-5.6 Sol Max.
Price and distribution
List price starts at $2 per million input tokens and $6 per million output tokens. A fast variant is twice that. Independent write-ups of the docs say rates double at or above 200,000 prompt tokens. Besides the SpaceXAI API, the launch named OpenRouter, Vercel, and Cloudflare. There is no open-weights release.
- First flagship launched under the SpaceXAI name
- Default in Grok Build; on all Cursor plans at launch
- No self-hosting path
Takeaways
- Timed with the Cursor close, not a coincidence in the launch copy
- The stated bet is multi-step agents, not a bigger base checkpoint
- Pricing undercuts several frontier list prices at the $2 / $6 headline
Source: SpaceXAI
