SpaceXAI has launched Grok 4.7, its latest flagship AI model targeted at software engineering, agentic automation, and general knowledge tasks. Built on a larger base architecture with extended reinforcement learning, the model maintains the pricing of Grok 4.6 at $2 per million input tokens and $6 per million output tokens. Grok 4.7 is immediately accessible via the xAI API, Cursor, Grok Build, OpenRouter, and major serverless platforms.

Benchmark performance shows significant increases over its predecessor, with Terminal-Bench 4.0 scores climbing from 20.3% to 38.0% and legal evaluation scores on Harvey’s benchmark reaching 19.6%. Additionally, SpaceXAI implemented a new safety stack, noting a 62.4% score on LatchBio’s biosafety evaluations and introducing a regional US endpoint for compliant enterprise inference.

By holding token prices flat while scaling capabilities, SpaceXAI is pressuring enterprise LLM pricing across coding and agent ecosystems. The launch highlights the rapid cadence of flagship model updates required to stay competitive against frontier rivals like Anthropic and OpenAI.

Why it matters

  • Price-performance pressure: Freezing API costs at $2/$6 while improving performance forces competing LLM vendors to lower prices.

  • Developer integration: Immediate default availability in popular coding environments like Cursor accelerates developer adoption.

  • Enterprise compliance: Dedicated regional API endpoints and revised safety stacks lower adoption barriers for enterprise and government clients.

Source: marktechpost.com