Submit
Moonshot AIOpen-weight·1049K context·— params·Kimi K3 License

Kimi K3moonshot/kimi-k3

Kimi K3 is Moonshot AI's frontier model, released 2026-07-16. At 2.8 trillion total parameters (104B activated) it is the largest open-weight model published to date, with a 1,048,576-token context window, native visual understanding, and an always-on reasoning mode. Weights ship under the custom Kimi K3 License with MXFP4 quantization-aware training. Pricing is the highest of any Chinese API in this directory: cache misses cost $3.00 per 1M input tokens, but cache hits drop to $0.30, so prompt-caching discipline changes the economics by 10x.

Pricing across providers

No current verified API pricing is listed. See the official model documentation for availability and billing.

Capabilities

  • chat
  • code
  • reasoning
  • long_context
  • agent
  • vision

Technical specs

Context
1049K
Max output
—
Parameters
—
Release
—
Training cutoff
—
License
Kimi K3 License
Hugging Face →

Similar models

Compare with

  • Kimi K3 vs DeepSeek V3.1
    Comparison planned — not yet published
  • Kimi K3 vs R1
    Comparison planned — not yet published
  • Kimi K3 vs DeepSeek V3
    Comparison planned — not yet published

Frequently asked

How much does Kimi K3 cost?+
No current verified API pricing is listed. This does not mean the model is free or that a hosted API is currently available. Check the official model documentation below.
How do I access Kimi K3 from outside China?+
Availability, account verification and payment methods vary by provider and region. Check the official documentation and the selected provider's access requirements before integrating. A directory listing does not guarantee global API access.
Is Kimi K3 open-source? Can I fine-tune it?+
Kimi K3 has published weights. License: Kimi K3 License. Review the linked repository and license for deployment, fine-tuning and commercial-use conditions.
What's the maximum context window for Kimi K3?+
The model supports up to 1,048,576 tokens of context (input + output). Some hosted versions may impose a smaller limit — check the "Context" column in the pricing matrix for each provider.