SpaceXAI Releases Grok 4.7 for Coding Agents and Long-Running Knowledge Work

SpaceXAI released Grok 4.7 on September 21, positioning the model as its strongest system for coding, agentic work and professional knowledge tasks, with API access, Cursor support and AI Gateway integrations available at launch. The update pairs a larger base model and longer reinforcement-learning run with lower token prices than several named rivals, while outside testing highlights a tradeoff: strong intelligence scores and low first-token latency, but slower sustained output than peer reasoning models.