Grok 4.7
Rank of 27093from #6
Known for
Long coding and agent jobs where the bill matters as much as the score.
SpaceXAI's newest Grok, out on 21 September 2026. A bigger base model than 4.6, trained on jobs that take hours, at the same $2 / $6 price. It leads its own launch table on electrical engineering and legal work.
What it does9
- 500K-token context
- Four reasoning levels, low to xhigh
- Text and image in, text out
- Knowledge cutoff May 2026
- $2 / $6 per million tokens, doubling above 200K
- Cached input at $0.50 per million
- Fast variant at twice the speed and twice the price, in Cursor and Grok Build only
- Web search, X search, code execution and function calling built in
- Keeps its own reasoning across turns on the Responses API
From Grok 4.7’s own site. Not used for ranking.
63index score · LLMs
How well does it score?
Compared with everyone else on this board.
—votes · crowd
Reading votes…
Votes don’t change the score.