supermarket

Language

Kimi K2.6

By Moonshot AI, in the Kimi K family.

Released
20 Apr 2026
Status
Legacy
Context
262.1K
Max output
—
Input per 1M tokens
$0.95
Output per 1M tokens
$4.00
Open weights
Yes
License
modified-mit
Parameters
1000B, 32B active
Knowledge cutoff
—
Reasoning
Yes
Modalities
text, image, video → text

Category scores

Out of 100, from independent and third-party results only, by method category-2026-09-29.1.

Category scores
CategoryScoreRankCoverage
Knowledge & reasoning22.8743 of 4
Factuality & hallucination34.9431 of 1
Human preference60.353 of 5

Benchmark results

Every result we hold for this model, with who ran it and where it was published.

Knowledge & reasoning

Knowledge & reasoning results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
Chess PuzzlesPuzzles solved26.0%Independent—7 May 2026Epoch AI
DTBench (Decision Theory Benchmark)Accuracy (capability questions)90.9%Third party——https://conceptualreasoning.ai/dtbench via Epoch AI
Earthborne Rangers (EBR-bench)Objectives reached2.4%Independent—25 Jun 2026Epoch AI
Humanity's Last Exam (HLE)Accuracy54.0%Self-reported——Model Card via Hugging Face Community Evals
LMCA (Language Model Conceptual Argumentation)Correlation with expert ratings (Pearson r)0.3726Third party——https://conceptualreasoning.ai/lmca via Epoch AI
Mystery Game PuzzlesPuzzles solved18.0%Independent—17 Jul 2026Epoch AI
Mystery Game PuzzlesPuzzles solved12.0%Independentnone29 Aug 2026Epoch AI

Scientific reasoning

Scientific reasoning results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
GPQA DiamondAccuracy90.8%Independent—1 May 2026Epoch AI
GPQA DiamondAccuracy90.5%Self-reported——Model Card via Hugging Face Community Evals

Mathematics

Mathematics results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
FrontierMath Tier 4Accuracy25.6%Independent—10 Jun 2026Epoch AI
FrontierMath Tier 4 (v1)Accuracy14.6%Independent—8 May 2026Epoch AI
FrontierMath Tiers 1-3Accuracy57.2%Independent—10 Jun 2026Epoch AI
FrontierMath Tiers 1-3 (v1)Accuracy39.0%Independent—7 May 2026Epoch AI
OTIS Mock AIME 2024-2025Accuracy96.1%Independent—2 May 2026Epoch AI

Coding & software engineering

Coding & software engineering results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
ExploitBenchMean capability18.4%Source not stated——Epoch AI via Epoch AI
ExploitBenchMean capability16.3%Source not stated——Epoch AI via Epoch AI
SWE-bench Verified% resolved76.7%Independent—8 May 2026Epoch AI
WeirdML (v2)Average accuracy55.9%Third party——https://htihle.github.io/weirdml.html via Epoch AI

Computer & browser use

Computer & browser use results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
OSWorld 2.0Binary completion (500 steps)4.6%Third party——https://osworld-v2.xlang.ai/ via Epoch AI

Vision & multimodal understanding

Vision & multimodal understanding results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
Furniture AssemblyScore21.7%Independent—11 Sep 2026Epoch AI

Factuality & hallucination

Factuality & hallucination results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
SimpleQA VerifiedAccuracy (% correct)34.9%Independent—10 Aug 2026Epoch AI

Human preference

Human preference results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
Code Arena: WebDev (arena.ai)Arena score (Elo scale)1509Independent—25 Sep 2026Arena
Code Arena: WebDev (arena.ai)Votes10202Independent—25 Sep 2026Arena
Document Arena (arena.ai)Arena score (Elo scale)1450Independent—13 Sep 2026Arena
Document Arena (arena.ai)Votes11291Independent—13 Sep 2026Arena
Text Arena (arena.ai)Arena score (Elo scale)1461Independent—25 Sep 2026Arena
Text Arena (arena.ai)Votes39971Independent—25 Sep 2026Arena
Vision Arena (arena.ai)Arena score (Elo scale)1265Independent—28 Sep 2026Arena
Vision Arena (arena.ai)Votes16416Independent—28 Sep 2026Arena

Other Kimi K versions

Sources