The LLM benchmark aggregator.

Kimi K2

Moonshot · 1T · open weights

Strongest in Tool use (#9 of 81), weakest in Toxicity avoidance (#244 of 272). Above par in 13 of 26 scopes. Among the models it meets almost everywhere, it finishes behind Claude Opus 5.5 and Claude Fable 5 and ahead of GPT OSS 120B and Qwen3.7 Max.

§ 1 · Profile

What it is good at

Bars run from 50 — the average of the models each source lists — so right of the line is above par. The middle column is the gap to whoever leads that scope.

Agents58.5−9.6#43/2681/5

Tool use65.4−8.6#9/811/1

Coding53.1−16#47/1651/5

Agentic coding53.8−14.9#50/1571/4

Professional49.1−12.5#91/1681/1

Legal52.5−14.6#59/1511/1

Medical50.2−16.6#79/1401/1

Finance46−16.9#109/1511/1

Knowledge48.1−15.4#98/1381/2

Academic knowledge47.7−16.2#90/1231/1

Core abilities49.8−17.6#100/2041/3

General intelligence49.7−20.4#101/2041/3

Human preference58.8−8.7#103/3421/1

Reasoning45−21.9#129/1782/4

Mathematics49.1−16.6#84/1401/2

Science46.8−16.7#87/1221/1

Reasoning39.3−29.8#115/1401/3

Safety51.9−9.6#135/3373/3

Safe-prompt compliance59.1−1.3#11/821/1

Harm refusal57.4−4.2#28/3002/2

Secure code63.2−3.3#30/2741/1

Fairness54.3−15.7#73/3002/2

Factual grounding30−40.6#92/1011/1

Jailbreak resistance37.7−27.2#228/2721/1

Toxicity avoidance37.3−21.7#244/2721/1

§ 2 · Head to head

What it beats, and what beats it

The same models turn up scope after scope. Counted once: where both were placed, who finished higher. Bars run right for Kimi K2, left for the other.

Gemini 2.0 Flashwon 21 · lost 2

Gemini 1.5 Flashwon 15 · lost 2

Mistral Medium 3.5won 14 · lost 2

GPT-4o Miniwon 20 · lost 3

GPT OSS 20Bwon 18 · lost 3

Mistral Smallwon 18 · lost 3

Gemini 1.5 Prowon 16 · lost 3

GPT-4.1 Nanowon 21 · lost 4

Command R Pluswon 15 · lost 3

Llama 4 Scout Instructwon 20 · lost 4

Command Awon 20 · lost 4

Mistral Largewon 14 · lost 5

Claude 3.5 Haikuwon 14 · lost 5

Mistral Large 3won 14 · lost 5

GPT-4.1 Miniwon 18 · lost 7

Gemini 2.5 Flash Litewon 14 · lost 6

Llama 4 Maverick Instructwon 18 · lost 8

Ling 3.0 Flashwon 11 · lost 6

DeepSeek V3won 16 · lost 9

GPT-4owon 14 · lost 8

GPT-5.4 Nanowon 12 · lost 7

O3 Miniwon 13 · lost 9

Gemini 3.5 Flash Litewon 10 · lost 7

Grok 4.3won 13 · lost 10

GPT-5.4 Miniwon 10 · lost 8

Claude 3.5 Sonnetwon 11 · lost 9

MiniMax M3won 13 · lost 11

Kimi K2.6won 13 · lost 11

Claude Haiku 4.5won 14 · lost 12

GPT-4.1won 14 · lost 12

Qwen3.6 Pluswon 9 · lost 8

Gemma 4 31Bwon 11 · lost 10

Qwen3.7 Maxwon 12 · lost 11

GPT OSS 120Bwon 13 · lost 12

MiniMax M2.7won 10 · lost 10

Mimo V2.5 Prowon 11 · lost 11

DeepSeek V3.2won 12 · lost 13

Grok 4.1 Fastwon 11 · lost 12

Gemini 3.5 Flashwon 11 · lost 12

Gemini 3.6 Flashwon 11 · lost 12

Grok 3won 10 · lost 11

GPT-5 Nanowon 11 · lost 13

GLM-5.1won 9 · lost 12

O4 Miniwon 11 · lost 15

Gemini 2.5 Flashwon 11 · lost 15

Grok 4won 10 · lost 15

GLM-5.3 Flashwon 9 · lost 14

Gemini 3.1 Flash Litewon 7 · lost 11

GLM-4.6won 9 · lost 15

Claude Sonnet 4won 9 · lost 16

Grok 4 Fastwon 6 · lost 12

Inklingwon 8 · lost 16

Qwen3 235B A22B Instructwon 8 · lost 17

Claude 3.7 Sonnetwon 7 · lost 15

Claude Sonnet 4.6won 7 · lost 15

GLM-4.5won 6 · lost 13

MiniMax M2.5won 5 · lost 11

DeepSeek V4 Flashwon 7 · lost 16

DeepSeek R1won 7 · lost 17

Kimi K2.5won 7 · lost 17

GPT-5.6 Lunawon 7 · lost 17

GLM-5.2won 7 · lost 17

O1won 6 · lost 15

Gemini 2.5 Prowon 7 · lost 18

GPT-5 Miniwon 7 · lost 19

Gemini 3.1 Prowon 5 · lost 14

Gemini 3 Prowon 6 · lost 17

Gemini 3 Flashwon 4 · lost 12

GLM-5won 4 · lost 14

Kimi K2 Thinkingwon 4 · lost 16

GPT-5.2won 5 · lost 20

Claude Sonnet 4.5won 5 · lost 21

GPT-6 Solwon 4 · lost 17

Claude Opus 4.6won 4 · lost 17

Grok 4.5won 4 · lost 19

Qwen3.8 27Bwon 3 · lost 15

Mimo V2.6 Prowon 3 · lost 15

Claude Opus 4won 4 · lost 20

DeepSeek V4 Prowon 4 · lost 20

GPT-5.4won 4 · lost 20

GPT-5won 4 · lost 21

GPT-6 Lunawon 3 · lost 17

Claude Opus 4.5won 3 · lost 17

GPT-5.1won 3 · lost 20

Muse Spark 1.1won 3 · lost 20

Gemini 3.8 Flashwon 3 · lost 20

Claude Sonnet 5won 3 · lost 21

GPT-5.6 Solwon 3 · lost 21

O3won 3 · lost 22

Qwen3.8 Maxwon 2 · lost 21

GPT-5.6 Terrawon 2 · lost 22

GPT-5.5won 2 · lost 22

Claude Opus 4.7won 2 · lost 22

DeepSeek V4.1 Flashwon 1 · lost 15

GPT-6 Astrawon 1 · lost 15

GLM-5.3won 1 · lost 16

Gemini 3.7 Flashwon 1 · lost 16

Grok 4.7won 1 · lost 19

Grok 4.6won 1 · lost 22

Claude Opus 4.8won 1 · lost 22

Claude Opus 5won 1 · lost 22

Muse Spark 1.2won 0 · lost 16

Kimi K3won 0 · lost 17

Claude Fable 5.1won 0 · lost 17

Claude Fable 5won 0 · lost 17

Claude Opus 5.5won 0 · lost 20

§ 3 · Sources

Where the numbers come from

9 publications, 24 figures. Every one links to the page it was read from.

LMArena Text ↗1 measureread 2026-09-26

LMArena Text 1419

Aider polyglot ↗1 measureread 2026-09-26

Aider polyglot 59.1%

BFCL ↗1 measureread 2026-09-26

BFCL v4 59.06%

Kagi LLM Benchmark ↗1 measureread 2026-09-26

Kagi LLM Benchmark 52.2%

SimpleBench ↗1 measureread 2026-09-26

SimpleBench 26.3%

Vals.ai ↗7 measuresread 2026-09-26

Vals · LegalBench 81.45%Vals · CorpFin 50.39%Vals · TaxEval 70.2%Vals · MedQA 83.97%Vals · GPQA Diamond 71.46%Vals · MMLU Pro 79.39%Vals · AIME 62.71%

HELM Safety ↗5 measuresread 2026-09-26

HELM Safety · HarmBench 97.4%HELM Safety · SimpleSafetyTests 100%HELM Safety · Anthropic Red Team 99.3%HELM Safety · BBQ 94.9%HELM Safety · XSTest 98.2%

Enkrypt AI Safety Leaderboard ↗6 measuresread 2026-09-26

Enkrypt · Jailbreak risk 23.2%Enkrypt · Harmful content risk 15.6%Enkrypt · CBRN risk 12.8%Enkrypt · Toxicity risk 15.2%Enkrypt · Bias risk 74.9%Enkrypt · Insecure code risk 6.7%

Vectara Hallucination Leaderboard ↗1 measureread 2026-09-26

Vectara · Factual consistency 82.1%

Badge

[![PublicAI Index](https://publicai.io/model-index/badge?model=kimi-k2)](https://publicai.io/model-index/m/kimi-k2)