ModelsSep 22, 2026 · 10:118 sources
SpaceXAI Launches Grok 4.7 With 19.6% Legal Benchmark Score to Beat GPT 5.6
From huggingnews.com
SpaceXAI launched Grok 4.7, scoring 19.6% on the Harvey Legal Agent Benchmark, outperforming GPT-5.6 and Fable 5.1, and reaching 38.0% on Terminal-Bench 4.0. The model is available via Grok Build, Cursor, and API, with pricing at $2 per 1 million input tokens and $6 per 1 million output tokens. Early user reports indicate pricing volatility and restrictive usage limits on the $300 per month supergrok heavy plan, with costs doubling for tasks exceeding 200,000 tokens or when using high-speed mode. Elon Musk recommended using the company's Build harness for optimal results with the model's updated reasoning and self-verification capabilities.
Sources
+ 6 more
- SUPPORT@rohanpaul_ai“On the Harvey Legal Agent Benchmark, Grok 4.7 scored 19.6%, while GPT-5.6 Sol Max scored 2.5% and Fable 5.1 Max scored 6.7%”x.com
- SUPPORT@theo“Comes out more expensive than Astra on Artificial Analysis”x.com
- SUPPORT@davis7“Over 200K tokens it doubles and with fast mode it increases even more”x.com
- SUPPORT@mark_k“Pricing starts at $2 per million input tokens and $6 per million output tokens”x.com
- SUPPORT@scaling01“being an Opus / Sol class model (~2T)”x.com
- SUPPORT@artificialanlys“Grok 4.7 scores 46 on the Artificial Analysis Intelligence Index to bring SpaceXAI into the top 4 AI labs”x.com
One ranked edition, every morning
The stories that matter today. Free, cancel anytime.
Telegram @aidr_today