Tag: Mixture of Experts
-
Liquid AI Releases LFM2.5-24B-A2B: A 2B-Active Open MoE That Runs Local Agents at 184 Tok/s
Liquid AI released LFM2.5-24B-A2B, an open-weight 24B Mixture-of-Experts model activating 2.45B parameters that hits 184 tok/s on an Apple M5 Max laptop.
-
DeepSeek-V3 and R1 MoE Architecture: How Sparse Routing Slashes Cloud Bills by 80%
DeepSeek-V3’s 671B parameter Mixture-of-Experts architecture activates only 37B parameters per token. Here is why sparse routing changes enterprise AI unit economics forever.