Artificial Analysis
#4of 110
Score 43.6. Best run: moonshotai/kimi-k3.
Intelligence Index: a composite of reasoning, knowledge, maths and coding evaluations.
Moonshot AI
Kimi K3 benchmarks, ranks and price
Where Artificial Analysis, LiveBench, LMArena and Berkeley BFCL place Kimi K3, what it costs through OpenRouter and Vercel AI Gateway, and the models either side of it on the consensus.
Read Oct 16, 2026
#5of 121
on the LLM leaderboard, with a consensus score of 82 across 3 leaderboards.
Source: Artificial Analysis, LiveBench, LMArena and Berkeley BFCL, combined by Rank.ai.
By leaderboard
Where each leaderboard places Kimi K3
LiveBench
#6of 36
Score 78.5. Best run: Kimi K3.
Contamination-limited questions refreshed each release: reasoning, maths, coding, data analysis, language.
LMArena
#7of 111
Score 1486. Best run: kimi-k3.
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
$3.13
OpenRouter: $0.30 in, $12 out. Vercel AI Gateway: $3.00 in, $15 out Context window 1.0M tokens. LiveBench cost per solved task $0.379.
Nearby
The models either side of Kimi K3
- #2GPT-5.6 Sol84
- #3Claude Opus 583
- #4GPT-5.582
- #5Kimi K382
- #6Claude Opus 4.881
- #7gemini-3-pro80
- #8Claude Opus 4.780
Questions about Kimi K3
How good is Kimi K3?
Kimi K3 ranks #5 of 121 models on the Rank.ai LLM leaderboard, with a consensus score of 82 out of 100 across 3 leaderboards.
Is Kimi K3 better than GPT-5.5?
Not on the consensus: GPT-5.5 scores 82 to Kimi K3's 82. Compare them source by source below, since the leaderboards test different things.
Is Kimi K3 better than Claude Opus 4.8?
On the consensus, yes: 82 to 81.
How much does Kimi K3 cost?
$0.30 per million input tokens and $12 per million output tokens on OpenRouter, the cheaper listing we read. Blended at three input tokens to one output, that is $3.13 per million.
Is Kimi K3 good for coding?
It scores 76.2 on the Artificial Analysis Coding Index, placing it #5 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.