SpaceXAI shipped Grok 4.7, and independent benchmarks place it fourth among frontier AI labs. According to BeInCrypto, Artificial Analysis scored the model at 46 on its Intelligence Index, 2 points above Grok 4.6, while Claude Fable 5.1, GPT-6 Astra and Claude Opus 5 ranked higher.

Elon Musk said before launch that Grok 4.7 would run on a 2.1-trillion-parameter base and include SpaceX company data. The model kept the same pricing as Grok 4.6, but Artificial Analysis said it uses far more tokens per task, which could lift costs.