Artificial Analysis
#40of 115
Score 23.5. Best run: moonshotai/kimi-k2.5.
Intelligence Index: a composite of reasoning, knowledge, maths and coding evaluations.
Moonshot AI
Kimi K2.5 benchmarks, ranks and price
Where Artificial Analysis, LiveBench, LMArena and Berkeley BFCL place Kimi K2.5, what it costs through OpenRouter and Vercel AI Gateway, and the models either side of it on the consensus.
Read Oct 5, 2026
#32of 126
on the LLM leaderboard, with a consensus score of 63 across 2 leaderboards.
Source: Artificial Analysis, LiveBench, LMArena and Berkeley BFCL, combined by Rank.ai.
By leaderboard
Where each leaderboard places Kimi K2.5
LiveBench
Not measured by LiveBench.
Contamination-limited questions refreshed each release: reasoning, maths, coding, data analysis, language.
LMArena
#31of 115
Score 1450. Best run: kimi-k2.5-thinking.
Blind head-to-head human votes, style controlled.
Berkeley BFCL
Not measured by Berkeley BFCL.
Function calling and tool use, single turn to multi turn and web search.
Price and context
$0.90
OpenRouter: $0.45 in, $2.25 out. Vercel AI Gateway: $0.60 in, $3.00 out Context window 262K tokens.
Nearby
The models either side of Kimi K2.5
- #29Qwen3.7 Plus66
- #30Claude Sonnet 4.665
- #31GPT-5.164
- #32Kimi K2.563
- #33DeepSeek V4 Pro 042363
- #34GLM 4.661
- #35Claude Opus 4.161
Questions about Kimi K2.5
How good is Kimi K2.5?
Kimi K2.5 ranks #32 of 126 models on the Rank.ai LLM leaderboard, with a consensus score of 63 out of 100 across 2 leaderboards.
Is Kimi K2.5 better than GPT-5.1?
Not on the consensus: GPT-5.1 scores 64 to Kimi K2.5's 63. Compare them source by source below, since the leaderboards test different things.
Is Kimi K2.5 better than DeepSeek V4 Pro 0423?
On the consensus, yes: 63 to 63.
How much does Kimi K2.5 cost?
$0.45 per million input tokens and $2.25 per million output tokens on OpenRouter, the cheaper listing we read. Blended at three input tokens to one output, that is $0.90 per million.
Is Kimi K2.5 good for coding?
It scores 46.8 on the Artificial Analysis Coding Index, placing it #36 for coding among the models on this leaderboard.
These models recommend brands. See if they name yours.
Enter your website. In about two minutes, Rank.ai asks ChatGPT, Claude and Gemini 12 questions your customers ask and grades how often they name you.
- Your grade out of 100How often AI names you, cites your site, and how high it ranks you.
- Who gets namedEvery competitor in the answers, most named first.
- The pages AI readsThe sources behind each answer.
- Three fixesWhat to fix first, with a brief for the first article.