**Moonshot** released the **Kimi K3**, a **2.8T-parameter MoE** model with **104B active parameters/token**, featuring innovations like **Kimi Delta Attention (KDA)**, **Gated MLA**, and **LatentMoE**. The release includes infrastructure co…
Firehose
Filtered to Companies, tagged “MoE models” · clear filters
Browse: People · Companies · Papers · Podcasts · Hacker News · Deep dives