MiniMax-M2.5
MiniMax-M2.5 is an open-weights AI model from MiniMax with an index score of 144 (GPT-4 = 100), rank 84 of 217.
144128–160Index score#84 of 217 models
$0.29/ $1.2Price per million tokensinput / output · via OpenRouter
198KContext windowtokens it reads at once
12 Feb 2026ReleasedReplaced by MiniMax-M2.7
Where it sitson the frontier
o3 reached this score 10 months earlier.
Test resultsby area · tick = best by any model
Reasoning & knowledgeevidence
Mathsno evidence yet
Codingevidence
Agentsno evidence yet
Professional workevidence
Novel problemsevidence
Long tasksno evidence yet
Third-party results only, at the model’s best standard setting. Hover a bar for who ran it.
What others sayrank on 8 leaderboards
The Model Indexthird-party tests, our method#84of 217 · 144
LMArena · Creative writingcreative writing#137of 360 · 1356LMArena · Codingcoding questions in chat#136of 357 · 1447LMArena · Long promptslong, complex requests#134of 340 · 1406LMArena · Instruction followingfollowing detailed instructions#143of 362 · 1381LMArena · Textoverall chat quality, judged by people#144of 362 · 1391LMArena · Mathsmaths questions in chat#142of 346 · 1392LMArena · WebDevbuilding web apps and front ends#95of 133 · 1386Best attop three among current models
Replaced by MiniMax-M2.7; it no longer leads any category among current models.
Where to use itprices checked 18 Nov 2026
API providers 4 · per 1M tokens, in / out
Run it yourself
Official weightsMiniMaxAI/MiniMax-M2.5↗Host prices from OpenRouter; going direct can differ.
Facts and sources
Released12 Feb 2026
Announcementhuggingface.co ↗
Release entryMiniMax-M2.5
Scored asMiniMax-M2.5
Output speed–
Hugging FaceMiniMaxAI/MiniMax-M2.5 ↗