SpaceXAI releases Grok 4.7 with larger base model and improved benchmarks

SpaceXAI has released Grok 4.7, a new flagship model for coding, agentic tasks, and knowledge work that utilizes a larger base model and extended reinforcement learning run compared to Grok 4.6. The model maintains the same pricing structure as its predecessor while offering enhanced capabilities in self-verification, long-context handling, and safety.

• Grok 4.7 features a new base model not reused from Grok 4.6 and a longer RL run weighted toward complex tasks.
• It achieves top scores on EEBench (64.0%) and the Harvey Legal Agent Benchmark (19.6%), with significant improvements on Terminal-Bench 4.0 (38.0%).
• The model includes a new safeguard stack, allowing only 3.3% of risky dual-use prompts through on HackerBench v0.3.
• Pricing remains $2 per million input tokens and $6 per million output tokens, with a 500,000-token context window and support for text/image input.

The release provides developers with a cost-effective option for professional knowledge work and coding tasks, available via the xAI API, Cursor, Grok Build, OpenRouter, Vercel, and Cloudflare.