New page: Grok 4.7 — frontier model for complex coding and knowledge work, with effort levels (xhigh, high, medium, low), 256k standard / 500k long context, tool access, and on-demand pricing (including Fast and long-context rates).
Changelog
Models
New page: Grok 4.7 — frontier model for complex coding and knowledge work, with effort levels (xhigh, high, medium, low), 256k standard / 500k long context, tool access, and on-demand pricing (including Fast and long-context rates).
Grok 4.6 and Grok 4.5 now recommend Grok 4.7 for stronger self-verification and longer-running work; both remain available.
Composer 2.5 docs note it shares the Cursor Models pool with Grok 4.7, Grok 4.6, and Grok 4.5.
Pricing and usage
Models and pricing: Grok 4.7 is included in the Cursor Models pool (with Grok 4.6, Grok 4.5, and Composer 2.5). Start (India) covers these models in non-fast mode with fixed medium effort for Grok 4.7/4.6/4.5.
Teams pricing: paid seats’ Cursor Models pool now lists Grok 4.7 alongside Grok 4.6, Grok 4.5, and Composer 2.5.
Evals / SDK
Run Cursor in your evals: default eval model is Grok 4.7 (grok-4.7); examples and copy updated accordingly.
