Grok 4.7 arrives from SpaceXAI at the same price as the previous release

SpaceXAI has released Grok 4.7, its new flagship model for coding, agentic tasks and knowledge work. The company says it is built on a larger base model and a longer reinforcement learning run, yet it ships at the same price and speed as Grok 4.6: $2 per million input tokens and $6 per million output tokens.

The model can be called today through the xAI API, Cursor, Grok Build, OpenRouter, Vercel and Cloudflare. It is available in Cursor on all plans and is the default model in Grok Build.

Benchmark numbers

All scores are vendor-reported. In SpaceXAI's table, Grok 4.7 improves on Grok 4.6 in every row. The biggest jump is Terminal-Bench 4.0, from 20.3% to 38.0%. EEBench rose 11 points to 64.0%, the highest figure in the table. On Harvey's legal agent benchmark, Grok 4.7 scored 19.6% against 6.7% for Fable 5.1 Max.

It does not lead everywhere. Fable 5.1 Max tops 4 of 7 benchmarks, including 57.9% on Terminal-Bench, and GPT-5.6 Sol Max holds DeepSWE v1.1 at 72.7%. On GDPval, which tests professional knowledge work, Grok 4.7 xhigh scored 1,695 Elo, up from 1,605 for Grok 4.6 high; Fable 5.1 max leads with 1,735 and GPT-6 Astra max scored 1,542.

Pricing is the other axis: Fable 5.1 Max costs 5x more on input and roughly 8.3x more on output, while GPT-5.6 Sol Max costs 2x and about 3.3x more.

Safeguards

SpaceXAI describes a new safeguard stack and calls Grok 4.7 the strongest model it has tested on refusals and jailbreak resistance. It topped LatchBio's biosafety benchmark at 62.4%, and on the company's own HackerBench v0.3, 3.3% of risky dual-use prompts got through.

Grok 4.7 Fast runs the same model on faster infrastructure at twice the output speed and twice the price, limited to Cursor and Grok Build. A US regional endpoint keeps inference in the country at a 10% premium.

Sources: SpaceXAI announcement.