AI TOOLS RADAR
All models / MOONSHOTAI
moonshotai logo
moonshotai

MoonshotAI: Kimi K3

Source description (English)

Kimi K3 is a 2.8T parameter open-weight multimodal reasoning model from Moonshot AI. It is suited for complex coding, knowledge work, and long-horizon agentic workflows, and is particularly strong at...

View source on OpenRouter Hugging Face
Overview

Model specifications

Context
1,048,576 tokens
Maximum output
943,718 tokens
Architecture
text+image+video->text
Tokenizer
Other
Knowledge cutoff
Not provided
Moderated
No
OPENROUTER

Complete pricing

Synchronized OpenRouter rates. Token prices are shown per one million tokens.

Input
$0.191
per 1M tokens
Output
$20
per 1M tokens
Cache read
$0.181
per 1M tokens
API

Model configuration

Default parameters

top_p
0.95

Reasoning

Reasoning required
No
Default parameters
max
Capabilities
max, high, low, none
API

Quick start

Set OPENROUTER_API_KEY locally. Python requires requests; JavaScript runs in Node.js. Keep the key on the server.

API documentation

curl --fail-with-body https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"moonshotai/kimi-k3","messages":[{"role":"user","content":"Hello!"}]}'
Capabilities

Input → Output

Input
textimagevideo
Output
text
Reasoning

Yes · max · high · low · none

API

Supported API parameters

frequency_penaltyinclude_reasoninglogit_biaslogprobsmax_completion_tokensmax_tokensmin_pparallel_tool_callspresence_penaltyreasoningreasoning_effortrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_ktop_logprobstop_p
Available providers

29 providers

Checked: October 26, 2026

Live providers on OpenRouter

Provider availability, latency, throughput and routing can change continuously. Open the source page for current operational data.

OpenRouter

Relace

fp4
Context
1.0M
Maximum output
944K
Input
$0.1814
Output
$8.5226
Cache read
$0.18
Cache write
Not provided

Wafer

mxfp4
Context
1.0M
Maximum output
944K
Input
$0.191
Output
$16
Cache read
$0.181
Cache write
Not provided

InferenceNet

fp4
Context
1.0M
Maximum output
944K
Input
$0.239
Output
$10.8
Cache read
$0.21
Cache write
Not provided

InferenceNet

fp4
Context
1.0M
Maximum output
944K
Input
$0.239
Output
$11.936
Cache read
$0.21
Cache write
Not provided

Wafer

mxfp4
Context
1.0M
Maximum output
944K
Input
$0.36
Output
$11.9
Cache read
$0.359
Cache write
Not provided

InferenceNet

fp4
Context
1.0M
Maximum output
944K
Input
$0.5
Output
$11.88
Cache read
$0.35
Cache write
Not provided

Makora

fp4
Context
1.0M
Maximum output
944K
Input
$0.65
Output
$15
Cache read
$0.23
Cache write
Not provided

Morph

fp8
Context
1.0M
Maximum output
944K
Input
$1.15
Output
$16.092
Cache read
$0.409
Cache write
Not provided

Sail Research

fp4
Context
1.0M
Maximum output
944K
Input
$1.2
Output
$5.5
Cache read
$0.2
Cache write
Not provided

Sail Research

fp4
Context
1.0M
Maximum output
944K
Input
$1.2
Output
$5.5
Cache read
$0.2
Cache write
Not provided

AkashML

fp4
Context
1.0M
Maximum output
944K
Input
$1.2
Output
$14
Cache read
$1.2
Cache write
Not provided

Decart

mxfp4
Context
1.0M
Maximum output
944K
Input
$1.65
Output
$8.25
Cache read
$0.165
Cache write
Not provided

Chutes

mxfp4
Context
1.0M
Maximum output
66K
Input
$1.85
Output
$9.5
Cache read
$0.185
Cache write
Not provided

Fireworks

Not provided
Context
1.0M
Maximum output
944K
Input
$2.1
Output
$10.5
Cache read
$0.21
Cache write
Not provided

InferenceNet

fp4
Context
1.0M
Maximum output
944K
Input
$2.24
Output
$14.519
Cache read
$0.35
Cache write
Not provided

Phala

Not provided
Context
1.0M
Maximum output
944K
Input
$2.55
Output
$12.75
Cache read
$0.255
Cache write
Not provided

DigitalOcean

Not provided
Context
1.0M
Maximum output
944K
Input
$2.55
Output
$12.95
Cache read
$0.255
Cache write
Not provided

DeepInfra

mxfp4
Context
1.0M
Maximum output
944K
Input
$2.85
Output
$14.25
Cache read
$0.285
Cache write
Not provided

ParetoAlpha

Not provided
Context
1.0M
Maximum output
944K
Input
$3
Output
$15
Cache read
$0.3
Cache write
Not provided

Amazon Bedrock

Not provided
Context
1.0M
Maximum output
128K
Input
$3
Output
$15
Cache read
$0.3
Cache write
$3.75

Parasail

fp4
Context
1.0M
Maximum output
944K
Input
$3
Output
$15
Cache read
$0.3
Cache write
Not provided

Modal

mxfp4
Context
1.0M
Maximum output
944K
Input
$3
Output
$15
Cache read
$0.3
Cache write
Not provided

Together

Not provided
Context
1.0M
Maximum output
944K
Input
$3
Output
$15
Cache read
$0.3
Cache write
Not provided

Fireworks

Not provided
Context
1.0M
Maximum output
944K
Input
$3
Output
$15
Cache read
$0.3
Cache write
Not provided

BaseTen

fp8
Context
1.0M
Maximum output
262K
Input
$3
Output
$15
Cache read
$0.3
Cache write
Not provided

Moonshot AI

mxfp4
Context
1.0M
Maximum output
944K
Input
$3
Output
$15
Cache read
$0.3
Cache write
Not provided

Alibaba

Not provided
Context
1.0M
Maximum output
131K
Input
$3.45
Output
$17.25
Cache read
$0.345
Cache write
Not provided

Fireworks

Not provided
Context
1.0M
Maximum output
944K
Input
$4.5
Output
$22.5
Cache read
$0.45
Cache write
Not provided

Fireworks

Not provided
Context
1.0M
Maximum output
944K
Input
$4.5
Output
$22.5
Cache read
$0.45
Cache write
Not provided
moonshotai

4 models

All models
moonshotai logomoonshotai

MoonshotAI: Kimi K3.1 Pro

Kimi K3.1 Pro is Moonshot AI's flagship open-weight multimodal reasoning model and a post-trained successor to Kimi K3, with a 1M-token context window. It is built for long-horizon coding, 3D...

Context
1.0M
Input
text · image · video
Output
text
Input: $2Output: $10per 1M tokens
View model details →
moonshotai logomoonshotai

MoonshotAI: Kimi K3 (batch)

Kimi K3 is a 2.8T parameter open-weight multimodal reasoning model from Moonshot AI. It is suited for complex coding, knowledge work, and long-horizon agentic workflows, and is particularly strong at...

Context
1.0M
Input
text · image · video
Output
text
Input: $2.28Output: $11.4per 1M tokens
View model details →
moonshotai logomoonshotai

MoonshotAI: Kimi K2.7 Code

MoonshotAI: Kimi K2.7 Code is a coding-focused model in Moonshot AI's Kimi K2 family, built to complete end-to-end programming tasks reliably over long contexts. It uses a native multimodal mixture-of-experts...

Context
262K
Input
text · image
Output
text
Input: $0.6712Output: $3.35per 1M tokens
View model details →
moonshotai logomoonshotai

MoonshotAI: Kimi K2.6

Kimi K2.6 is Moonshot AI's next-generation multimodal model, designed for long-horizon coding, coding-driven UI/UX generation, and multi-agent orchestration. It handles complex end-to-end coding tasks across Python, Rust, and Go, and...

Context
262K
Input
text · image
Output
text
Input: $0.65Output: $3.41per 1M tokens
View model details →