The Model Index
Up to four models

Kimi K2.6 vs Kimi K2.7 Code

On the index (GPT-4 = 100), Kimi K2.6 scores 152 and Kimi K2.7 Code 151. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.

Next to Kimi K2.6
Presets
Kimi K2.6
152136–169
WeightsOpen
ReleasedApr 2026
Price in / out$0.65 / $3.41
Context262K
Core tests taken4 of 9
Kimi K2.7 Code
151135–167
WeightsOpen
ReleasedJun 2026
Price in / out$0.70 / $3.75
Context262K
Core tests taken3 of 9

Kimi K2.6 and Kimi K2.7 Code are too close to call (2 points apart, within the uncertainty), ahead on 2 of 2 shared core tests, at a similar price.

Test by testthird-party results

Kimi K2.6Kimi K2.7 Code
0%25%50%75%100%REASONING & KNOWLEDGEGPQA Diamond (Epoch AI)GPQA Diamond (Epoch AI)LiveBench Data AnalysisLiveBench Data AnalysisLiveBench IFLiveBench IFLiveBench LanguageLiveBench LanguageLiveBench ReasoningLiveBench ReasoningSimpleQA VerifiedSimpleQA VerifiedMATHSFrontierMath (tiers 1-3)FrontierMath (tiers 1-3)AIME-style maths (OTIS mock)AIME-style maths (OTIS mock)LiveBench MathematicsLiveBench MathematicsChess puzzlesChess puzzlesFrontierMath (tier 4)FrontierMath (tier 4)CODINGLiveBench Agentic CodingLiveBench Agentic CodingLiveBench CodingLiveBench CodingCode MigrationCode MigrationLiveCodeBenchLiveCodeBenchProgramBenchProgramBenchSWE-benchSWE-benchVibe Code Bench v1.1Vibe Code Bench v1.1WeirdMLWeirdML
Every result as a table
TestKimi K2.6Kimi K2.7 Code
Reasoning & knowledge
GPQA Diamond (Epoch AI)90.8%87.9%
LiveBench Data Analysis65.1%62.7%
LiveBench IF64.4%56.3%
LiveBench Language75.1%77.9%
LiveBench Reasoning79.4%82.8%
SimpleQA Verified34.9%36.5%
Maths
FrontierMath (tiers 1-3)57.2%54.0%
AIME-style maths (OTIS mock)96.1%95.6%
LiveBench Mathematics84.3%79.6%
Chess puzzles20.0%14.0%
FrontierMath (tier 4)25.6%12.2%
Coding
LiveBench Agentic Coding46.9%45.7%
LiveBench Coding78.6%74.0%
Code Migration27.8%25.4%
LiveCodeBench86.8%82.0%
ProgramBench0.0%0.0%
SWE-bench76.2%78.2%
Vibe Code Bench v1.137.9%47.2%
WeirdML55.9%54.1%

Tests at least two of these models have taken. Bold rows are the core tests.

This page as a chartDownload PNG