Try Kimi

Model selection

Kimi API provides a variety of models for developers to choose from. Different models have different strengths in capability, speed, and price.

Main models

kimi-k3

kimi-k3 is Kimi's flagship model, built for long-horizon coding and end-to-end knowledge work, with native visual understanding. K3 always runs in thinking mode. You can set the thinking effort via the top-level reasoning_effort parameter (supports low, high, max; defaults to max). It supports a context window of up to 1M tokens.

kimi-k2.6

kimi-k2.6 supports text, image, and video input, and supports switching between thinking and non-thinking modes. It is suitable for conversation, code generation, visual understanding, and agent tasks. It supports a context window of up to 256k tokens.

For the full list of available models and their parameters, visit platform.kimi.ai/docs/introduction.

Selection dimensions

When choosing a model, evaluate the following dimensions:

  • Context length: Different models support different maximum context windows. Choose a larger-context model for long-document processing.
  • Response speed: Lighter models respond faster, making them suitable for latency-sensitive scenarios.
  • Generation quality: Higher-tier models perform better on complex reasoning and creative tasks.
  • Price: Choose the most cost-effective model based on your budget and call volume.
  • Thinking mode: Choose kimi-k3 for deep reasoning, or kimi-k2.6 when you need flexibility to switch thinking mode on and off.

Vision models (image understanding)

Vision models support image input and can be used for image description, OCR, chart interpretation, and similar tasks:

  • Each image is billed at a fixed 1024 tokens, regardless of image size or resolution.
  • Common image formats are supported (JPEG, PNG, WebP, etc.).
  • Images can be passed via URL or Base64 encoding.

Currently unsupported capabilities

  • PPT generation API: PPT generation is not currently available via API.
  • Deep research API: Deep research is not currently available via API.

Watch the platform announcements for the latest updates.

Model selection - Kimi Help Center