Grok 4
Grok 4 is a closed AI model from xAI with an index score of 148 (GPT-4 = 100), rank 71 of 217.
148132–164Index score#71 of 217 models
–Price per million tokensno public API price found
–Context windowtokens it reads at once
9 Jul 2025ReleasedReplaced by Grok 4.1
Where it sitson the frontier
Among the first models to reach this score.
Test resultsby area · tick = best by any model
Reasoning & knowledgeevidence
Mathsevidence
Codingevidence
Agentsevidence
Professional workevidence
Novel problemsevidence
Long tasksevidence
Third-party results only, at the model’s best standard setting. Hover a bar for who ran it.
What others sayrank on 10 leaderboards
The Model Indexthird-party tests, our method#71of 217 · 148
LMArena · Creative writingcreative writing#99of 360 · 1394LMArena · Mathsmaths questions in chat#98of 346 · 1425LMArena · Textoverall chat quality, judged by people#124of 362 · 1411LMArena · Instruction followingfollowing detailed instructions#130of 362 · 1396Epoch Capabilities Indexoverall capability from many benchmarks#84of 231 · 147.0LMArena · Long promptslong, complex requests#124of 340 · 1416LMArena · Codingcoding questions in chat#145of 357 · 1435LMArena · Visionunderstanding images#89of 145 · 1184LMArena · Searchanswering with web search#35of 41 · 1142Best attop three among current models
Replaced by Grok 4.1; it no longer leads any category among current models.
Where to use itprices checked 18 Nov 2026
Host prices from OpenRouter; going direct can differ.
Facts and sources
Released9 Jul 2025
Announcementx.ai ↗
Release entryGrok 4 (+ Heavy)
Scored asGrok 4
Output speed–
Hugging Faceclosed weights