xAI launches Grok 4.7 with low pricing but weaker benchmark results
xAI has released Grok 4.7, presenting it as its most capable model so far for coding and knowledge work. The company says the system uses a larger base model, longer reinforcement-learning training and mechanisms intended to improve verification of its own outputs. Access is priced at $2 per million input tokens and $6 per million output tokens.
On the independent Artificial Analysis Intelligence Index (v4.3.2), which combines ten benchmarks, Grok 4.7 received 46 points and ranked in the middle of the measured models. Claude Fable 5.1 and GPT-6 each scored 53. The difference was larger in agentic coding: Grok 4.7 achieved 26 percent on Terminal-Bench 4.0, compared with 60 percent for GPT-6 Astra and 55 percent for Claude Fable 5.1. DeepSeek V4.1 Flash reached 27 percent. Grok 4.7 is available through the Grok API, Cursor and Grok Build.
Written by Sakhanda Wire AI from reporting by The Decoder. Facts, quotes and full context are in the original article below.