Kimi Chat API
Calling and billing for Kimi K2 Instruct and K2.5.
Use an official-group APIPool key with base URL https://app.apipool.dev/v1.
Check availability in the Kimi catalog
before calling. Models marked “Coming soon” do not accept requests yet.
Models
| Model ID | Model |
|---|---|
kimi-k2-instruct | Kimi K2 Instruct |
kimi-k2.5 | Kimi K2.5 |
Replace model in the example below with one of the IDs in the table.
Request
curl https://app.apipool.dev/v1/chat/completions \
-H "Authorization: Bearer $APIPOOL_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"kimi-k2.5","messages":[{"role":"user","content":"Hello"}],"stream":false}'Text chat uses POST /v1/chat/completions. For streaming, set stream: true
and stream_options: {"include_usage": true}. Read the complete stream for final usage.
Read non-streaming replies from choices[0].message.content. For streaming,
concatenate choices[].delta.content and keep reading until data: [DONE].
The returned usage contains the usage for this request.
Billing
Input and output tokens are billed separately; reasoning tokens count toward output. Cached input uses the regular input rate. See Kimi model details for current rates and the usage page for charges after a request.