Artificials
Compare models
All makers

AI model maker

Moonshot AI

17 models on Artificials, 8 with published test results and 14 with open weights you can download.

RSS feed
Best on the capability index157.4Kimi K3, #15 of 274
Best rankTop 2%Kimi K3 (max) on Arena hard prompts
Newest releasekimi-for-codingSep 11, 2026
Models178 tested, 14 with open weights

Where Moonshot AI stands

Its best model’s rank among all models tested on each skill area’s main test, the same tests model pages use. A longer bar means a better rank.

  • OverallKimi K3 on Epoch Capabilities Index
    Top 6%
  • CodingKimi K3 on LiveBench coding
    Top 21%
  • MathKimi K3 (max) on Mock AIME 2024–2025
    Top 14%
  • KnowledgeKimi K3 (max) on SimpleQA Verified
    Top 32%
  • AgentsKimi K3 on LiveBench agentic coding
    Top 15%
  • LanguageKimi K3 on LiveBench language
    Top 16%
  • Data analysisKimi K3 on LiveBench data analysis
    Top 31%
  • ReasoningKimi K3 (max) on GPQA Diamond
    Top 9%

On the capability index

Its 6 models on the Epoch Capabilities Index, each with its rank among all 274 models. The top score is 167.3, held by Claude Opus 5.5 from Anthropic.

Every test

Its best result on each published test, and that result’s rank among every model tested. Scores from different tests are never combined.

TestBest modelScoreRank
Epoch Capabilities IndexKimi K3157.4#15 of 274
Arena hard promptsKimi K3 (max)1517#7 of 389
Arena codingKimi K3 (max)1541#7 of 384
Arena instruction followingKimi K3 (max)1488#11 of 389
Arena textKimi K3 (max)1488#13 of 389
Arena mathKimi K3 (max)1501#14 of 372
Arena creative writingKimi K3 (max)1460#20 of 387
GPQA DiamondKimi K3 (max)93.1%#19 of 222
Arena WebDevKimi K3 (max)1658#11 of 123
LiveBench reasoningKimi K390.7%#8 of 63
Mock AIME 2024–2025Kimi K3 (max)97.2%#26 of 198
LiveBench agentic codingKimi K362.2%#9 of 63
LiveBench languageKimi K385.5%#10 of 63
Chess puzzlesKimi K3 (max)39.0%#24 of 141
LiveBench codingKimi K381.5%#13 of 63
LiveBench overallKimi K379.2%#13 of 63
FrontierMath Tiers 1–3Kimi K3 (max)72.2%#21 of 81
SWE-bench Verified, bash onlyKimi K2.5 (high)70.8%#11 of 42
SWE-bench VerifiedKimi K2.676.7%#9 of 32
LiveBench data analysisKimi K378.7%#19 of 63
DeepSWEKimi K3 (max)68.5%#8 of 26
SimpleQA VerifiedKimi K3 (max)50.6%#24 of 77
FrontierMath Tier 4Kimi K3 (max)39.0%#22 of 63
LiveBench instructionsKimi K371.4%#23 of 63
LiveBench mathematicsKimi K384.4%#50 of 63

Every model

Newest first. The price is per 1M tokens, blended three input tokens to one output token: Moonshot AI’s own listing, or else the middle price across hosts.

ModelReleasedContextPriceCapability index
kimi-for-codingSep 11, 20261M tokensNo paid listingNot on the index
Kimi K3Jul 16, 20261M tokens$6.00157.4
Kimi K3 FastJul 16, 20261M tokens$9.00Not on the index
Kimi K2.7 CodeJun 12, 2026262.1K tokens$1.71150.0
Kimi For Coding HighSpeedJun 12, 2026262.1K tokensNo paid listingNot on the index
Kimi K2.7 Code HighSpeedJun 12, 2026262.1K tokens$3.43Not on the index
Kimi K2.6Apr 21, 2026262.1K tokens$1.71151.1
Kimi LatestApr 21, 20261M tokens$3.87Not on the index
Kimi K2.5Jan 2026262.1K tokens$1.20148.0
Kimi K2 ThinkingNov 6, 2025262.1K tokens$1.08Not on the index
Kimi K2 Thinking TurboNov 6, 2025262K tokens$2.86Not on the index
kimi-k2-0905-previewSep 5, 2025262.1K tokens$1.11Not on the index
Kimi K2 0905Sep 4, 2025262.1K tokens$1.08Not on the index
Kimi K2Dec 1, 2024131.1K tokens$1.00Not on the index
Kimi K2 ThinkingNov 1, 2024262.1K tokens$1.08146.0

Models, prices, context windows and release dates from models.dev (MIT). Test results from LiveBench (Apache-2.0), Epoch AI (CC BY 4.0), Datacurve DeepSWE, compiled by Epoch AI (CC BY 4.0), Arena (CC BY 4.0) and SWE-bench (CC BY-NC 4.0). A model counts here when Moonshot AI sells it, several hosts list it or a test covers it. Each figure is the source’s own; nothing is combined into a new score.