Vision / Moonshot
Kimi-K3
API name: kimi-k3
Visual UnderstandingDeep Thinking
Overview
Kimi-K3 is Moonshot AI’s most capable flagship model to date, featuring 2.8 trillion parameters. Built on the KDA hybrid linear attention mechanism (Kimi Delta Attention) and Attention Residuals, it natively supports visual understanding and offers a 1-million-token context window. As the world's first open-source model at the 3-trillion-parameter scale, it is designed for frontier intelligence scenarios such as long-horizon programming, knowledge work, and reasoning.
- Model type: Vision
- Input: Text / Image / Video
- Output: Text
- Endpoints: chat / chat_completion
Tiered billing details
| Tier | Input | Output | Cache hit | Cache write |
|---|---|---|---|---|
| Fallback tier | $0.0030/1K | $0.0150/1K | $0.0003/1K | $0.0030/1K |
Billed bylen (input context token count). Coefficient is the $ / 1M tokens price. Default group price (other groups converted by discount).
Capabilities
- Programming: Long-range coding capabilities, adept at tasks that integrate software engineering with visual reasoning.
- Knowledge Work: The agent’s knowledge-work capabilities have been comprehensively enhanced.
Code example
Call with kimi-k3 :
curl https://www.starunion.net/v1/chat/completions \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "kimi-k3",
"messages": [{
"role": "user",
"content": [
{ "type": "text", "text": "Describe this image" },
{ "type": "image_url", "image_url": { "url": "https://example.com/photo.jpg" } }
]
}]
}'