InfoReleasesHacker News · community
Kimi K3.1 Pro and K3.1 Pro Highspeed
Kimi Code has fully rolled out K3.1 Pro, its most capable 3T-class model with a 1M context window, plus K3.1 Pro Highspeed offering roughly 2–3× the speed with identical capabilities. The K3.1 Pro models only support thinking mode (requests with thinking disabled are rejected), and the API supports OpenAI Responses/Chat Completions and Anthropic protocols. Switching models or thinking effort mid-session invalidates the context cache and increases token usage, and mistyped model IDs silently fall back to K2.8 Preview without thinking.
Read the full story on kimi.com
What to doUse exact model IDs like k3.1-pro or k3.1-pro-highspeed, keep thinking enabled, and start new sessions when switching models or effort levels.
- Source
- kimi.com
Sources
- Hacker News · communityKimi K3.1 Pro and K3.1 Pro Highspeed
- Hacker News · communityKimi K3.1 Pro and K3.1 Pro Highspeed