2 tracked versions of the Kimi K family. Timeline ordered newest → oldest. Flagship version marked.
Kimi K2.7 Code is Moonshot AI's coding-agent specialist, released 2026-06-12. A 1-trillion-parameter Mixture-of-Experts model (32B active, 384 experts) with a 256K-token context window, it targets long-horizon agentic software engineering and uses roughly 30% fewer reasoning tokens than K2.6 while scoring higher on Moonshot's coding benchmarks. Open weights under a Modified MIT license.
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.
See every provider hosting Moonshot AI models at /provider/moonshot.