Kimi-K3 API pricing: 17 providers compared
17 providers serve Kimi-K3. The cheapest charges $1.50 per million input tokens, the most expensive $6.00. Put in your monthly usage to see what that gap costs you.
| Provider | $/M in | $/M out | $/M cached in | Your month | Precision | Context | Uptime 1d | Tok/s |
|---|
Shipping with AI? New models, SDK releases and price changes that matter, checked against the sources. One email a day.
Precision is what each provider reports. Some models ship in fp8 or fp4, so a low number alone does not mean a cut-down copy. If a provider runs lower precision than the release, its answers can differ from the version you tested. "unknown" means the provider does not say.
About Kimi-K3
Moonshot AI drops the Kimi K3 open weights on Hugging Face, alongside the K3 technical report. K3 is a 2.8T-parameter Mixture-of-Experts model (104B active) with a 1M-token context, Delta Attention hybrid, and native vision — the model is now self-hostable in addition to the kimi.com app and OpenAI-compatible API.
What changed in this release · Model card on Hugging Face
Where these prices come from
Prices, precision, context, uptime and throughput are read live from OpenRouter's public endpoint list for moonshotai/kimi-k3 each time this page loads, so they match what the providers charge today. If that request fails, the page shows the snapshot taken when it was built, and says so. Prices are per million tokens in US dollars, before any provider discounts or free tiers.