Revana
Sign in Get API key
Language models/Kimi/Kimi K3
Kimitext2 routes

Kimi K3

Kimi's most capable model, with visual understanding.

kimi-k3Get API key

Pricing USD per million tokens

TokensPer 1M tokens
Input$1.80
Cached input$0.18
Cache write (5 min)$1.80
Cache write (1 hour)$3.60
Output$9.00

Reasoning tokens are billed as output. Each answer's usage carries cost_usd, the amount taken from your balance; a request with no answer costs $0.00.

Request

curl
curl https://api.revana.si/v1/chat/completions \
  -H "Authorization: Bearer $REVANA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"kimi-k3","messages":[{"role":"user","content":"Write a haiku about routing."}]}'
Python
import os
from openai import OpenAI

client = OpenAI(base_url="https://api.revana.si/v1", api_key=os.environ["REVANA_API_KEY"])
reply = client.chat.completions.create(
    model="kimi-k3",
    messages=[{"role": "user", "content": "Write a haiku about routing."}],
)
print(reply.choices[0].message.content)

Any OpenAI SDK works: set the base URL to https://api.revana.si/v1. Streaming, tools and the rest are in the chat completions docs.

Details

Model id
kimi-k3
Maker
Kimi
Endpoint
POST /v1/chat/completions
Context window
1M tokens
Max output
1M tokens
Streaming
Yes (server-sent events)
Routes
2 (falls back automatically)

2 routes

Every request starts on the cheapest route. A busy or failed route hands it to the next before any answer is sent.

About the Kimi K3 API

How much does the Kimi K3 API cost?

Kimi K3 costs $1.80 per million input tokens, $0.18 per million cached input tokens and $9.00 per million output tokens on Revana. You pay from prepaid credit for the tokens each request used, with no subscription or minimum.

How do I call Kimi K3 from code?

Use any OpenAI SDK with base URL https://api.revana.si/v1 and your Revana API key, and set model to "kimi-k3". POST /v1/chat/completions takes the OpenAI request body, streamed or not.

What happens if a request fails?

Revana has 2 routes for Kimi K3 and moves a request to the next one when a route is busy or down. A request that gets no answer costs nothing.

Is Kimi K3 on Revana the same model?

Yes. Revana passes your request to Kimi's model unchanged and returns its answer in the OpenAI format, with the usage it reported and what the request cost.

More from Kimi

Also as markdown for agents. Every language model and price: llms.json.