Elon Musk’s xAI has released Grok 4.7, positioning it as its most capable model for coding and knowledge tasks. The company states the system uses a larger base model and extended reinforcement learning to improve output verification. Pricing is set at $2 per million input tokens and $6 per million output tokens, rates that align more closely with Chinese models than Western frontier offerings. Independent testing on the Artificial Analysis Intelligence Index places Grok 4.7 at a score of 46, putting it in the middle of the pack. Claude Fable 5.1 and GPT-6 both achieve a score of 53, establishing a clear performance lead.
The disparity widens significantly when measuring agentic coding capabilities. On Terminal-Bench 4.0, Grok 4.7 scores just 26 percent compared to 60 percent for GPT-6 Astra and 55 percent for Claude Fable 5.1. Even the cheaper DeepSeek V4.1 Flash model edges ahead with a 27 percent score. The new model is available through the Grok API, Cursor, and Grok Build.
* Scores lag behind leaders by 7 points on general benchmarks
* Agentic coding performance trails competitors by roughly 34 points
* Pricing strategy targets cost-sensitive markets rather than premium tiers
