Wiredesk
Models·Hugging Face·17:05 UTCSwap-ready

Run fast, private agents on-device with LFM2.5-24B-A2B

Read the source →
Why it matters

Gives developers high-throughput, low-memory local inference options for agentic extraction and catalog preprocessing tasks.

Summary

Liquid AI released LFM2.5-24B-A2B, an open non-transformer architecture optimized for fast on-device agent execution.