Skip to main content
Moonshot AIJul 16, 2026

Kimi K3

kimi-k3PrivateOpen weights

Kimi K3 is an ultra-large-scale, 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 navigating large repositories, using tools, debugging, and iterating against images, logs, tests, and runtime feedback.

$3.75/$18.75
per 1M in / out
1M
context window
128K
max output
Variant
Also available end-to-end encrypted$3.75 / $18.75 per 1M
How it comparesamong 143 text models
GPQA Diamondhigher than 78%91.9%
Blended pricecheaper than 15%$7.50
Context windowlarger than 42%1M
Max outputlarger than 78%128K

Further right is better. Hover a dot for its model.

Time to first token— 
Output speed— 
Uptime— 

Last 24 hours of Venice traffic.How it’s measured

Capabilities

What the model takes in, returns and supports through the API.

InTextImage
OutText

Not supportedvideo input, audio input, X search and log probabilities

Benchmarks

GPQA Diamond
91.9%29th of 139 · 14 score higher for less, including GPT-6.2 Sol, GPT-6.2 Sol Pro and Gemini 4 Argon

Specs

Model ID
kimi-k3
Endpoint
POST /api/v1/chat/completions
Context window
1,000,000 tokens
Max output
131,072 tokens
Reasoning
none · low · high · max · default max
Precision
Not disclosed
Privacy
Private
Added
Jul 16, 2026

Compare

ModelInputOutputGPQA DiamondContextMax outputFirst tokenSpeed
Kimi K3$3.75$18.7591.9%1M128K——
GPT-5.6 Sol$4.00$20.0093.4%1M128K——
GPT-5.6 Sol Pro$4.00$20.0094.9%1M128K——
Claude Opus 5.5$4.00$20.0093.4%1M128K——
Gemini 4 Neon$3.75$18.7595.5%1M64K——

Other versions

  1. Kimi K2.6Apr 2026
  2. Kimi K3Jul 2026

Pricing

Billed per token in USD. Reasoning tokens bill as output.

Input$3.75per 1M tokens
Cached input$0.37590% off input
Output$18.75per 1M tokens

Estimate a workload

$0.0169per request
$169for 10,000 requests

Blended price against similar models

Kimi K3$7.50

Per 1M tokens at three input tokens for each output token.

Quickstart

OpenAI-compatible: point any OpenAI SDK at the Venice base URL with a Venice API key.

curl https://api.venice.ai/api/v1/chat/completions \
-H "Authorization: Bearer $VENICE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "kimi-k3",
"messages": [{ "role": "user", "content": "Explain TEE attestation in two sentences." }],
"reasoning_effort": "max"
}'
Base URL https://api.venice.ai/api/v1

Parameters

ParameterValuesNotes
modelkimi-k3Required
messagestext · image_urlContent parts the model reads
max_completion_tokensUp to 131,072Counts reasoning tokens
reasoning_effortnone | low | high | maxDefault max
toolsFunction definitionsWith tool_choice
response_format{ type: "json_schema" }Output that matches a schema
streamtrue | falseServer-sent events
venice_parameters.enable_web_searchauto | on | offWeb results and citations