Kimi K2
Kimi K2 is an open-weights AI model from Moonshot with an index score of 131 (GPT-4 = 100), provisional.
131115–147Index scoreProvisional · results in fewer than three areas so far; not ranked yet
$0.57/ $2.3Price per million tokensinput / output · via OpenRouter
131KContext windowtokens it reads at once
11 Jul 2025ReleasedReplaced by Kimi K2 (Sep 2025)
Where it sitson the frontier
o1 reached this score 7 months earlier.
Test resultsby area · tick = best by any model
Reasoning & knowledgeevidence
Mathsno evidence yet
Codingevidence
Agentsno evidence yet
Professional workno evidence yet
Novel problemsno evidence yet
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 7 leaderboards
LMArena · Textoverall chat quality, judged by people#111of 362 · 1418LMArena · Codingcoding questions in chat#110of 357 · 1468LMArena · Creative writingcreative writing#115of 360 · 1383LMArena · Mathsmaths questions in chat#114of 346 · 1416LMArena · Instruction followingfollowing detailed instructions#133of 362 · 1390LMArena · Long promptslong, complex requests#133of 340 · 1406Epoch Capabilities Indexoverall capability from many benchmarks#92of 231 · 145.5
Best attop three among current models
Replaced by Kimi K2 (Sep 2025); it no longer leads any category among current models.
Where to use itprices checked 18 Nov 2026
API providers 1 · per 1M tokens, in / out
Run it yourself
Official weightsmoonshotai/Kimi-K2-Instruct↗Host prices from OpenRouter; going direct can differ.
Facts and sources
Released11 Jul 2025
Announcementkimi.com ↗
Release entryKimi K2
Scored asKimi K2 (Jul 2025)
Output speed–
Hugging Facemoonshotai/Kimi-K2-Instruct ↗