Kimi K3 reasoning effort tiers are now documented for Standard, High, and Max in Kimi Code — Moonshot AI's first meaningful update to its 2.8-trillion-parameter open-weight model since July 16. At launch only Max was live; the Kimi Code docs now map all three levels, with lighter tiers still flagged as coming soon on the main Kimi API.
Kimi K3 Adds Standard and High Reasoning Modes: Documentation Maps Three Effort Tiers
