Calculations happen locally. Token values are not sent anywhere.
Estimated spend
¥0.30
- Cached input
- ¥0.000000
- Standard input
- ¥0.20
- Output
- ¥0.10
- Per request
- ¥0.30
- All requests
- ¥0.30
Kimi K3 public API ratesCached ¥2 · Input ¥20 · Output ¥100 per MTokVerified 2026-07-19Official pricing source
Price configuration
Current public K3 rates
ModelCached inputInputOutputContext
Kimi K3¥2 / MTok¥20 / MTok¥100 / MTok1M tokens
How the estimate works
Each token type is divided by one million, multiplied by its official unit price, and then multiplied by your request count. Cached input only applies when the official platform recognizes a cache hit.
total = ((cached / 1M × ¥2) + (input / 1M × ¥20) + (output / 1M × ¥100)) × requests
Clear answers
Frequently asked questions
Is the Kimi K3 API free?
The official platform lists metered K3 API prices. Promotions or account credit may reduce your bill but do not make the API permanently free.
What is a cached input token?
It is input recognized by the provider's caching system and billed at a separate rate. Cache eligibility follows the official API behavior and cannot be assumed for every repeated prompt.
Does this estimator send my token counts anywhere?
No. It performs the arithmetic in your browser and does not submit the values to a server.