Grok 4.7

xAI has launched Grok 4.7, a larger base model utilizing extended reinforcement learning to enhance coding and knowledge work performance. The model achieves 46.3% on CursorBench 4.0 and 71% on DeepSWE 1.1, maintaining low refusal rates for legitimate use. While xAI describes it as a notable improvement over Grok 4.6 at the same price, benchmarks place it in second place. The update introduces multi-hour coding capabilities and $2 input pricing to accelerate productivity and innovation.