Models · Updated 23 Oct, 10:29 pm IST
Kimi launches K3.1 Pro and K3.1 Pro Highspeed with 1M-token context
Why it matters for readers: It shows progress in open-weight models offering long context windows and faster interaction options for complex tasks.
- K3.1 Pro is a post-trained upgrade to K3 with larger post‑training compute and optimizations across RL, long-context inference, and training infrastructure.1
- The model provides a 1 million token context window and is positioned to improve performance on coding, 3D creation, financial modeling, and agentic benchmarks.12
- K3.1 Pro Highspeed has the same capabilities but runs about 2–3× faster for more responsive interactions.12
- Kimi made K3.1 Pro and Highspeed available immediately via Kimi.ai, Kimi Work, Kimi Code and the Kimi API, with full model weights to be released by Nov 4.1
- Kimi Code lists multiple model IDs across the K3.1 Pro family and recommends using native APIs and starting new sessions for best results.2
Get a brief like this every morning
Uzha reads hundreds of sources and gives you the stories that matter for your work, with every source linked. Free.
Get startedHow this story developed
23 Oct, 10:29 pm · Hacker News (100+ points)
Kimi K3.1 Pro and K3.1 Pro Highspeed
23 Oct, 10:29 pm · Hacker News (100+ points)
Kimi K3.1 Pro and K3.1 Pro Highspeed