Kimi (Moonshot AI) API: endpoints, authentication, models and limits

Connection details

Base URLhttps://api.moonshot.ai/v1
AuthBearer token via MOONSHOT_API_KEY
OpenAI SDK compatibleYes
SDKsOpenAI SDK, Anthropic-compatible endpoint, raw HTTP

Models and token pricing

ModelModel IDContextIn / 1MOut / 1M
Kimi K3kimi-k31049K$3.00$15.00
Kimi K2.6kimi-k2.6256K$0.730$3.49

What is distinctive about this API

  • OpenAI-compatible chat completions, with an Anthropic-compatible surface also available — useful if your codebase is already built around Claude-style messages.
  • Long-context handling is the headline capability; the K-series models are built for large document and repository ingestion.
  • Strong agentic tool-use behaviour, and the open-weight releases can be self-hosted.

Constraints worth knowing first

  • Two regional endpoints exist (international and mainland China) with different hosts and separate keys — using the wrong pair returns an auth error that reads like an invalid key.
  • Very long contexts raise time-to-first-token substantially; measure prefill latency, not just tokens per second.

Frequently asked

What is the Kimi (Moonshot AI) API base URL?
https://api.moonshot.ai/v1. Authenticate with a bearer token, conventionally read from the MOONSHOT_API_KEY environment variable.
Is the Kimi (Moonshot AI) API compatible with the OpenAI SDK?
Yes. The chat completions surface is wire-compatible, so the official OpenAI client libraries work by pointing base_url at https://api.moonshot.ai/v1 and supplying a MOONSHOT_API_KEY. No client library change is needed.
How do I get a Kimi (Moonshot AI) API key?
Create an account with the provider and generate a key from the developer console, then expose it to your application as MOONSHOT_API_KEY. Keys are secrets: keep them server-side, never in client bundles or committed config.
Which models does the Kimi (Moonshot AI) API expose?
Kimi K3, Kimi K2.6. Context windows run to 1049K tokens.

Other provider APIs

Moonshot AI publishes official documentation at https://platform.moonshot.ai/docs. Pricing and limits change; verify against the provider before committing to a budget.

Deploy a model with Swfte Connect

One gateway, every provider, per-token cost visibility. Swap models without touching your code.