llmboard.ai
Benchmarks
CompareRankings
llmboard.ai
Benchmarks
CompareRankings
HomeModelsKimi K2.6

Moonshot AI model product

Kimi K2.6

Kimi K2.6 is Moonshot AI's open-source, native multimodal agentic model focused on state-of-the-art coding, long-horizon execution, and agent swarm capabilities. It scales horizontally to 300 sub-agents executing 4,000 coordinated steps, dynamically decomposing tasks into parallel, domain-specialized subtasks. K2.6 unifies text, image, and video input with thinking and non-thinking modes, supports a 256K context, and powers proactive 24/7 background agents that manage schedules, execute code, and orchestrate cross-platform operations without human oversight.

Updated Aug 10, 2026. Default version: Kimi K2.6

Compare
LLMBoard score82.8Kimi K2.6
Coverage100%8 benchmark families
Context window262.1KTokens
Official input price$0.95Moonshot AI API

On this page

  • Specification
  • Capability
  • Benchmarks
  • Arena
  • Pricing
  • Versions
  • About
  • Compare
  • Similar models
  • FAQ

Model specification

Structured fields from the published default version.

Version
Kimi K2.6
Released
Apr 20, 2026
Knowledge cutoff
Jan 1, 2025
Parameters
1T
Context window
262.1K
Max output
262.1K
Inputs
image, text, video
Outputs
text
Open weights
Yes
License
Modified MIT License

Capability profile

This profile uses the latest version under this unique model that has a calculated LLMBoard score. Arena and price are excluded.

Kimi K2.6 category scores

Benchmark results

Published benchmark records for the scored version Kimi K2.6.

28 rows
Columns

Show columns

Claw-Eval0.8113100.0%CAug 7, 2026
OJBench0.619100.0%CAug 7, 2026
V*1.017100.0%CAug 7, 2026
WideSearch0.819100.0%CAug 7, 2026
AIME 20261.021793.8%CAug 7, 2026
LiveCodeBench v60.925398.1%CAug 7, 2026
MathVision0.933293.5%CAug 7, 2026
MCP-Mark0.64857.1%CAug 7, 2026
APEX-Agents0.35733.3%CAug 7, 2026
BabyVision0.75950.0%CAug 7, 2026
HMMT Feb 260.951160.0%CAug 7, 2026
SciCode0.551876.5%CAug 7, 2026
DeepSearchQA0.86828.6%CAug 7, 2026
SWE-bench Multilingual0.863484.8%CAug 7, 2026
BrowseComp0.975889.5%CAug 7, 2026
IMO-AnswerBench0.971966.7%CAug 7, 2026
CharXiv-R0.984784.8%CAug 7, 2026
Finance Agent v20.4112660.0%BAug 7, 2026
SWE-Bench Verified0.81210489.3%CAug 7, 2026
FrontierSWE0.3131514.3%BAug 7, 2026
MMMU-Pro0.8146579.7%CAug 7, 2026
OSWorld-Verified0.7142238.1%CAug 7, 2026
Toolathlon0.5143156.7%CAug 7, 2026
Terminal-Bench 2.00.7154970.8%CAug 7, 2026
SWE-Bench Pro0.6174462.8%CAug 7, 2026
GPQA0.91923392.2%CAug 7, 2026
LiveBench0.7253835.1%BAug 7, 2026
Humanity's Last Exam0.4399258.2%CAug 7, 2026

Arena results

Preference and agent-evaluation signals from published Arena datasets.

100 rows
Columns

Show columns

textchinese121530.81,873N/AAug 6, 2026
text factualityjapanese151441.6342N/AAug 6, 2026
text style controlchinese151529.21,873N/AAug 6, 2026
text factualitychinese161520.21,601N/AAug 6, 2026
text factualitymath161472.51,625N/AAug 6, 2026
text factualityspanish161458.51,005N/AAug 6, 2026
text factualityindustry mathematical171475.11,670N/AAug 6, 2026
webdevimage to webdev171520.13,301N/AAug 4, 2026
text factualityfrench181487.91,196N/AAug 6, 2026
vision style controlchinese181329.8862N/AAug 6, 2026
visionchinese191349.1862N/AAug 6, 2026
visionhomework201310.02,149N/AAug 6, 2026
documentoverall211450.911,181N/AJul 30, 2026
textspanish211463.91,284N/AAug 6, 2026
visionhumor211280.3539N/AAug 6, 2026
vision style controlhomework221302.22,149N/AAug 6, 2026
vision style controlhumor221261.8539N/AAug 6, 2026
textexpert231489.93,376N/AAug 6, 2026
textjapanese231444.9455N/AAug 6, 2026
text style controlexpert241502.63,376N/AAug 6, 2026
text style controlmath241479.41,921N/AAug 6, 2026
visionenglish241285.26,414N/AAug 6, 2026
document style controloverall251449.511,181N/AJul 30, 2026
textkorean251426.5712N/AAug 6, 2026
textmath251474.11,921N/AAug 6, 2026
visionocr251290.911,072N/AAug 6, 2026
visionoverall261281.415,386N/AAug 6, 2026
visiondiagram261297.54,040N/AAug 6, 2026
textindustry mathematical271479.01,998N/AAug 6, 2026
text factualityexpert271497.52,835N/AAug 6, 2026
webdevoverall271509.79,261N/AAug 6, 2026
webdevwebdev271509.79,261N/AAug 6, 2026
webdevwebdev-react271505.16,764N/AAug 6, 2026
textindustry software and it services281483.914,915N/AAug 6, 2026
text style controlindustry mathematical281485.31,998N/AAug 6, 2026
visioncreative writing vision281270.1822N/AAug 6, 2026
vision style controlenglish281265.66,414N/AAug 6, 2026
textcoding291487.110,310N/AAug 6, 2026
text style controljapanese291446.0455N/AAug 6, 2026
vision style controloverall301263.415,386N/AAug 6, 2026
vision style controldiagram301291.74,040N/AAug 6, 2026
vision style controlocr311278.511,072N/AAug 6, 2026
textoverall321454.837,236N/AAug 6, 2026
textexclude ties321454.527,573N/AAug 6, 2026
textlonger query321466.215,197N/AAug 6, 2026
textnon english321446.820,191N/AAug 6, 2026
text factualityindustry software and it services321498.514,668N/AAug 6, 2026
text style controlcoding321515.010,310N/AAug 6, 2026
textenglish331458.217,045N/AAug 6, 2026
texthard prompts331469.424,032N/AAug 6, 2026
text style controlindustry software and it services331503.814,915N/AAug 6, 2026
vision style controlcreative writing vision331250.6822N/AAug 6, 2026
webdevwebdev-html331495.21,028N/AAug 6, 2026
text factualitycoding341513.99,774N/AAug 6, 2026
textindustry business and management and financial operations351445.77,391N/AAug 6, 2026
textfrench361468.61,491N/AAug 6, 2026
texthard prompts english361469.111,617N/AAug 6, 2026
text factualitylonger query361474.114,772N/AAug 6, 2026
text style controllonger query361475.915,197N/AAug 6, 2026
textinstruction following371448.811,994N/AAug 6, 2026
text factualityindustry business and management and financial operations371462.36,537N/AAug 6, 2026
text factualitynon english371442.820,163N/AAug 6, 2026
textgerman381450.1728N/AAug 6, 2026
textindustry life and physical and social science381467.46,152N/AAug 6, 2026
textindustry legal and government391458.82,807N/AAug 6, 2026
text style controlkorean391422.3712N/AAug 6, 2026
text style controlspanish391459.81,284N/AAug 6, 2026
text factualityoverall401455.237,190N/AAug 6, 2026
text factualityhard prompts401480.024,004N/AAug 6, 2026
text factualityhard prompts english401487.711,152N/AAug 6, 2026
text factualityindustry entertainment and sports and media401433.07,014N/AAug 6, 2026
text factualityindustry legal and government401468.02,287N/AAug 6, 2026
text factualityinstruction following401454.511,511N/AAug 6, 2026
textindustry entertainment and sports and media411425.27,541N/AAug 6, 2026
text factualityindustry life and physical and social science411477.35,337N/AAug 6, 2026
text factualityindustry medicine and healthcare411481.22,159N/AAug 6, 2026
text factualityexclude ties421457.227,537N/AAug 6, 2026
text factualityrussian421459.23,098N/AAug 6, 2026
text style controlhard prompts431485.024,032N/AAug 6, 2026
text style controlinstruction following431454.511,994N/AAug 6, 2026
text style controlnon english431449.520,191N/AAug 6, 2026
textindustry writing and literature and language441435.68,734N/AAug 6, 2026
text factualityenglish441463.517,027N/AAug 6, 2026
text style controloverall441460.737,236N/AAug 6, 2026
text style controlhard prompts english441486.911,617N/AAug 6, 2026
text style controlexclude ties451464.327,573N/AAug 6, 2026
textcreative writing461430.15,778N/AAug 6, 2026
text factualityindustry writing and literature and language461440.78,170N/AAug 6, 2026
text style controlgerman461457.7728N/AAug 6, 2026
textmulti turn481452.66,490N/AAug 6, 2026
textrussian481444.73,783N/AAug 6, 2026
text style controlenglish481467.617,045N/AAug 6, 2026
text factualitycreative writing491433.74,964N/AAug 6, 2026
textpolish511440.8737N/AAug 6, 2026
text factualitymulti turn521461.05,605N/AAug 6, 2026
text style controlfrench531475.51,491N/AAug 6, 2026
text style controlindustry life and physical and social science531477.66,152N/AAug 6, 2026
textindustry medicine and healthcare541453.32,638N/AAug 6, 2026
text style controlindustry business and management and financial operations541455.87,391N/AAug 6, 2026
text style controlcreative writing561429.55,778N/AAug 6, 2026

Pricing

Official vendor API PAYG pricing is summarized first. The table then lists individual provider offerings without treating their minimum as the official price.

Official API
$0.95 input, $4 output per 1M
Official provider
Moonshot AI
Lowest third-party
From $0.22 input, $1.14 output per 1M via LLM Gateway
Tracked offerings
23
22 rows
Columns

Show columns

Privatemode AIkimi-k2.6globalN/AN/A262.1KAug 7, 2026
Alibaba Token Plankimi-k2.6globalN/AN/A262.1KAug 7, 2026
Alibaba Token Plan (China)kimi-k2.6globalN/AN/A262.1KAug 7, 2026
LLM Gatewaykimi-k2.6global$0.22$1.14262.1KAug 7, 2026
routing.runkimi-k2.6global$0.275$1.1200KAug 7, 2026
CrofAIkimi-k2.6global$0.50$1.99262.1KAug 7, 2026
GreenPTkimi-k2.6global$0.7524$4.28262.1KAug 7, 2026
DigitalOceankimi-k2.6global$0.76$3.2262.1KAug 7, 2026
Cortecskimi-k2.6global$0.773$3.38262.1KAug 7, 2026
EBCloudKimi-K2.6global$0.9286$3.86262.1KAug 7, 2026
Alibaba (China)kimi-k2.6global$0.929$3.86262.1KAug 7, 2026
OpenCode Gokimi-k2.6global$0.95$4262.1KAug 7, 2026
OpenCode Zenkimi-k2.6global$0.95$4262.1KAug 7, 2026
Azure Cognitive Serviceskimi-k2.6global$0.95$4262.1KAug 7, 2026
Moonshot AI (China)kimi-k2.6global$0.95$4262.1KAug 7, 2026
Aurikokimi-k2.6global$0.95$4262.1KAug 7, 2026
Azurekimi-k2.6global$0.95$4262.1KAug 7, 2026
AIHubMixkimi-k2.6global$0.95$4262.1KAug 7, 2026
Charm Hyperkimi-k2.6global$0.95$4262KAug 7, 2026
Moonshot AIkimi-k2.6global$0.95$4262.1KAug 7, 2026
WaferKimi-K2.6global$1.14$4.8262.1KAug 7, 2026
UnoRouterkimi-k2.6global$1.27$5.34262.1KAug 7, 2026

Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.

Kimi K2.6 versions

All published versions linked to this unique model. The score columns identify the version used by the current overall ranking.

1 rows
Columns

Show columns

Kimi K2.6Apr 20, 202682.81T262.1K262.1KYesModified MIT License

What is Kimi K2.6?

A concise description based on the published model registry.

Kimi K2.6 is Moonshot AI's open-source, native multimodal agentic model focused on state-of-the-art coding, long-horizon execution, and agent swarm capabilities. It scales horizontally to 300 sub-agents executing 4,000 coordinated steps, dynamically decomposing tasks into parallel, domain-specialized subtasks. K2.6 unifies text, image, and video input with thinking and non-thinking modes, supports a 256K context, and powers proactive 24/7 background agents that manage schedules, execute code, and orchestrate cross-platform operations without human oversight.

Use the benchmark, Arena and pricing sections above as separate evidence. A missing field means the current data snapshot does not support that claim.

Data snapshot: 2026-08-07. Editorial model content is not available in the backend.

Kimi K2.6 vs nearby models

Open a comparison with the three ranked models immediately above and below this model.

Kimi K2.6vsGPT-5.6-TerraKimi K2.6vsGemini 3.1 ProKimi K2.6vsClaude Opus 4.8Kimi K2.6vsGemini 3 ProKimi K2.6vsSeed 2.1 TurboKimi K2.6vsGLM 5.2

Models similar to Kimi K2.6

Recommendations prioritize the same model type and family, then the closest published LLMBoard score.

#29-11.2
MA

Kimi K2.5

Moonshot AI

71.6 LLMBoard

DetailsCompare
#53-23.4
MA

Kimi K2 Thinking

Moonshot AI

59.4 LLMBoard

DetailsCompare
#80-31.3
MA

Kimi K2

Moonshot AI

51.4 LLMBoard

DetailsCompare
#6+0.1
AN

Claude Opus 4.8

Anthropic

82.9 LLMBoard

DetailsCompare
#8-0.5
GO

Gemini 3 Pro

Google

82.2 LLMBoard

DetailsCompare
#5+1.6
GO

Gemini 3.1 Pro

Google

84.3 LLMBoard

DetailsCompare

FAQ

Common questions about Kimi K2.6.

When was Kimi K2.6 released?

Kimi K2.6's default version was released on Apr 20, 2026.

How much does Kimi K2.6 cost?

Kimi K2.6's official API price is $0.95 per million input tokens and $4 per million output tokens via Moonshot AI. The lowest tracked third-party offer starts at $0.22 input and $1.14 output via LLM Gateway.

Who created Kimi K2.6?

Kimi K2.6 is published under Moonshot AI in the model registry.

What is the context window for Kimi K2.6?

The default version has a 262.1K token context window.

Is Kimi K2.6 open weight?

Yes. The default version is marked as open weight under Modified MIT License.

How many API providers offer Kimi K2.6?

23 published provider offerings are linked to the default version.

What models should I compare Kimi K2.6 with?

Nearby ranked alternatives include GPT-5.6-Terra, Gemini 3.1 Pro, Claude Opus 4.8.

Rankings

OverallCodingText ArenaPricing

Modalities

Image GenerationVideo GenerationSpeech-to-TextEmbeddings

Benchmarks

All BenchmarksReasoningMathCoding

Vendors

All VendorsOpenAIAnthropicGoogle
llmboard.aiCopyright 2026 llmboard.ai