Open frontier intelligence: 2.8T params, 1M-token context
Today, we are introducing Kimi K3 — our most capable model. Kimi K3 is a 2.8T-parameter model built on our Kimi Delta Attention and Attention Residuals, with native vision capabilities and a 1-million-token context window. It is the world's first open 3T-class model, designed for frontier intelligence across long-horizon coding, knowledge work, and reasoning.
No comments yet. Be the first!
Real conversations about Kimi K3 on X
Post on X
