Moonshot AIProvisional

Kimi K2.6

Kimi / Kimi K2.6 · moonshotai/Kimi-K2.6

Current mainstream Kimi open/API model for long-horizon coding and multimodal agent work.

Objective score

46.6

Confidence

86%

Price evidence

Pricing evidence pending

Support this model Compare trend Official source

Other primary tiers in this family

Kimi K2.6Kimi K2.7 Code

Scenario evidence

Six task-specific rankings

Feature development#20Obj. 46.60 · 0%Debug and repair#20Obj. 59.60 · 0%Refactoring#23Obj. 560 · 0%Code review#23Obj. 56.70 · 0%Large-repo agent work#19Obj. 43.10 · 0%Frontend / UI#25Obj. 54.70 · 0%

Evidence state

Signal coverage

Benchmarks: 3
Rank points: 7
Usage: 2
Agent: 0
Sources: 7

Objective evidence

Benchmarks and public signals

Swebench: 71.2
Livecodebench: —
Aider: —
Artificial Analysis: 61.8
Openrouter Usage: 0.5
Kilo Usage: —

Persistent history

Recent ranking points

Aug 1#15 · 45

Aug 1#15 · 45

Jul 31#15 · 45

Jul 31#15 · 45

Jul 31#15 · 45

Jul 31#15 · 45

Usage signals

OpenRouter0.5%

Overall0.3%

Agent and tool signals

Evidence is being collected. Missing data is not treated as zero.

Source trail

Official Moonshot Kimi K2.6 model card documents the current open model line.Artificial Analysis Coding Agents · ok WebDev Arena · ok Kilo leaderboard · ok Kimi K2.7 Code model card · ok NVIDIA Nemotron 3 Ultra model card · ok Kimi API models · ok Kimi K2.6 model card · ok