Spacexai launches grok 4.7 with enhanced capabilities at same pricing as previous version

SpaceXAI has launched Grok 4.7, a new flagship model designed for coding, agentic tasks, and knowledge work. The model features a larger base and extended reinforcement learning training, yet maintains the same pricing structure of $2 per million input tokens and $6 per million output tokens as its predecessor, Grok 4.6. It is now available through multiple platforms including the xAI API, Cursor, Grok Build, OpenRouter, Vercel, and Cloudflare. Grok 4.7 shows significant improvements across several benchmarks, including a 38.0% score on Terminal-Bench 4.0 compared to Grok 4.6's 20.3%. It also achieved a top score of 64.0% on EEBench and outperformed Fable 5.1 Max on Harvey’s legal agent benchmark with 19.6% versus 6.7%. Despite these gains, Fable 5.1 Max leads in four out of seven benchmarks, while GPT-5.6 Sol Max holds the highest DeepSWE v1.1 score at 72.7%. The model also includes a new safeguard stack, achieving a 62.4% score on LatchBio’s biosafety benchmark and allowing only 3.3% of risky prompts through on HackerBench v0.3.