Overview
A 1T-total, 32B-active open MoE focused on agentic tool use and coding.
Only published specifications are shown.
A 1T-total, 32B-active open MoE focused on agentic tool use and coding.
K2 paired trillion-scale MoE with an open agent model and demonstrated large-scale training with the Muon optimizer.
Kimi K2 launched as an open 1T MoE agent model.
View sourceSelect a card to view the related model.