DataLearner logo
KI

Kimi K2.6

Reasoning modelKimi K2Kimi K2.6

Kimi K2.6

Release date: 2026-04-20Updated: 2026-07-17 21:57:19.964Knowledge cutoff: 2025-0411,520
Live demoGitHubHugging FaceCompare
Parameters
1T
Context length
256K
Chinese support
Supported
Reasoning ability

Kimi K2.6 is a reasoning model from Moonshot AI, released on 2026-04-20. It accepts text, image, and video input and returns text output. The cataloged parameter count is 1T, with 32B active parameters per inference. The recorded context window is 256K. Cataloged capabilities include Reasoning model and Multilingual. The checkpoint is listed under the Modified Mit license, with the license link included in the model references. Use the linked references to confirm current access, licensing, and provider-specific limits.

Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology

Kimi K2.6

Model basics

Reasoning traces
Supported
Thinking modes
Thinking modes not supported
Context length
256K tokens
Max output length
No data
Model type
Reasoning model
Modality (in / out)
Text, Image, Video → Text
Release date
2026-04-20
Model file size
No data
MoE architecture
Yes
Total params / Active params
1T / 32B
Knowledge cutoff
2025-04
Kimi K2.6

Open source & experience

Code license
Modified MIT
Weights license
Modified MIT- Commercial use permitted
GitHub repo
GitHub link unavailable
Hugging Face
Hugging Face link unavailable
Live demo
https://www.kimi.com/
Kimi K2.6

Official resources

Paper
Paper
DataLearnerAI blog
No blog post yet
Kimi K2.6

API details

API speed
No data
💡Default unit: $/1M tokens. If vendors use other units, follow their published pricing.
Learn about pricing modes
Standard
TypeConditionInputOutput
Text-$0.950/ 1M$4.00/ 1M
Cache PricingPrompt Cache
TypeTTLWriteRead
Text1h$0.950/ 1M$0.160/ 1M
Kimi K2.6

Benchmark Results

Kimi K2.6 currently shows benchmark results led by LiveCodeBench (7 / 123, score 89.60), HLE (15 / 172, score 54), GPQA Diamond (18 / 187, score 90.50). This page also consolidates core specs, context limits, and API pricing so you can evaluate the model from benchmark results and deployment constraints together.

Thinking
Tool usage
Internet

General Knowledge

4 evaluations
Benchmark / mode
Score
Rank/total
GPQA Diamond
Thinking Mode
90.50
18 / 187
LiveBench
Thinking Mode
72.17
28 / 115
HLE
Thinking Mode
34.70
76 / 172
HLE
Thinking ModeToolsInternet
54
15 / 172

Coding and Software Engineer

4 evaluations
Benchmark / mode
Score
Rank/total
LiveCodeBench
Thinking Mode
89.60
7 / 123
SWE-bench Verified
Thinking ModeTools
80.20
14 / 112
SWE-bench Multilingual
Thinking ModeTools
76.70
5 / 23
SWE-Bench Pro - Public
Thinking ModeTools
58.60
13 / 54

AI Agent - Information Search

1 evaluations
Benchmark / mode
Score
Rank/total
BrowseComp
Thinking ModeToolsInternet
83.20
14 / 53

AI Agent - Tool Usage

4 evaluations
Benchmark / mode
Score
Rank/total
OSWorld-Verified
Thinking ModeTools
73.10
14 / 24
Terminal Bench 2.0
Thinking ModeTools
66.70
10 / 47
TerminalBench 2.1
Thinking Mode
53.56
27 / 27
Tool Decathlon
Thinking ModeTools
50
2 / 9

Math and Reasoning

2 evaluations
Benchmark / mode
Score
Rank/total
AIME 2026
Thinking Mode
96.40
3 / 18
IMO-AnswerBench
Thinking Mode
86
8 / 21

Claw-style Agent Evaluation

1 evaluations
Benchmark / mode
Score
Rank/total
Claw Bench
Thinking ModeTools
80.90
19 / 29
View benchmark analysisCompare with other models

Compare with other models

Kimi K2.6

Publisher

Kimi K2.6

Model Overview

Kimi K2.6 is a reasoning model from Moonshot AI, released on 2026-04-20.

It accepts text, image, and video input and produces text output. Its cataloged capabilities include Reasoning model and Multilingual. The cataloged parameter count is 1T, with 32B active parameters per inference. The recorded context window is 256K.

The checkpoint is listed under the Modified Mit license, with the license link included in the model references. The page records 4 API pricing rules from the listed provider; current provider pricing and conditions should be checked before deployment. The evaluation section contains 16 cataloged benchmark results with their recorded modes and scores. The page links 2 release, model-card, repository, or provider references for checking the underlying claims. Specifications, availability, and prices can change; undisclosed values are intentionally left unstated.

Kimi K2.6

FAQ

What is Kimi K2.6?

Kimi K2.6 is a reasoning model from Moonshot AI, released on 2026-04-20. It accepts text, image, and video input and returns text output. The cataloged parameter count is 1T, with 32B active parameters per inference. The recorded context window is 256K. Cataloged capabilities include Reasoning model and Multilingual. The checkpoint is listed under the Modified Mit license, with the license link included in the model references. Use the linked references to confirm current access, licensing, and provider-specific limits.

What input and output modalities does Kimi K2.6 support?

The current model record lists text, image, and video as input and text as output.

What are the main recorded specifications for Kimi K2.6?

The cataloged parameter count is 1T, with 32B active parameters per inference. The recorded context window is 256K. Fields without a source-backed value remain undisclosed.

Does Kimi K2.6 have API pricing?

The page records 4 API pricing rules from the listed provider; current provider pricing and conditions should be checked before deployment.

Are benchmark results available for Kimi K2.6?

The evaluation section contains 16 cataloged benchmark results with their recorded modes and scores. Compare only results that use the same benchmark version and evaluation mode.

Is Kimi K2.6 open source?

The checkpoint is listed under the Modified Mit license, with the license link included in the model references. Review the linked license text before commercial or derivative use.

DataLearner on WeChat

Follow DataLearner on WeChat for AI model updates and research notes.

DataLearner WeChat QR code