Introducing LFM3: Every Modality for Specialized Intelligence
Summary
Liquid AI introduces LFM3, a new generation of liquid foundation models designed for specialized intelligence. The 2B-parameter, multimodal family uses a tri-hybrid architecture to enable efficient on-device deployment across text, vision, audio, omni, and decision modalities, with a large context window and open-weight weights available for local fine-tuning. The release emphasizes edge deployment, performance benchmarks, and industry collaboration.