Grok 4.7 launches with coding gains but mixed frontier benchmarks

xAI/SpaceXAI's Grok 4.7 is available for coding, agents, and knowledge work through APIs and integrations including GitHub Copilot. It reports benchmark and safety gains over Grok 4.6 at unchanged low headline pricing, but high reasoning-token usage can weaken real-world ROI and independent evaluations vary substantially by effort and methodology.