Kimi API rates and the membership gap
The API platform is pay as you go. It is not the same bill as Kimi membership or Kimi Code. A membership monthly price was not in the docs we could cite.
Checked 2026-11-01. Rates below are USD per 1 million tokens. K3 context is 1,048,576 tokens, with no length tier.
K3 API rates
| Model | Input | Output | Cached input | 5-minute write | 1-hour write |
|---|---|---|---|---|---|
| kimi-k3.1-pro | $2.00 | $10.00 | $0.20 | $2.00 | $4.00 |
| kimi-k3 | $3.00 | $15.00 | $0.30 | $3.00 | $6.00 |
Batch for kimi-k3.1-pro and kimi-k3 is 60% of the standard price. The default cache TTL is 5 minutes. From 31 October 2026, K2 models are closed to newly registered users, and the docs plan to retire K2 by the end of December 2026. K2 rates are not in this table.
Source: Kimi API model pricing ↗
The docs say the API has no subscription and point at a product comparison, but that comparison did not include a citable monthly price. No number is added here.
Cross-country subscription samples
The table is this site’s price snapshot. Each row has a source and an observation date. These are historical App Store samples, not today’s checkout and not API rates. Rows are sorted by country name so you can look one up. That order is not a cheapest-country ranking.
Filter them in the matrix (中文) →Ofox next to the official API
Lock the same model, then compare the official input, output, and cache rates with the rates in the Ofox catalog. This page does not type in an Ofox price, and it does not mark any call as smoke-tested. An aggregator can make several models easier to manage. It is not a guaranteed discount, and it may not cover every feature on this page.