LFM3 is Liquid AI’s third generation of foundation models. Its first six roughly 2B-parameter models cover text, vision, audio, multimodal, and decision tasks for edge deployment; Liquid reports a 256K context for its 1.7B text model.
Features
- Open-weight small models
- Multimodal model family
- On-device deployment
Use Cases
- Evaluate mobile inference
- Build robotics model uses
FAQ
LFM3 is Liquid AI’s third generation of foundation models. Its first six roughly 2B-parameter models cover text, vision, audio, multimodal, and decision tasks for edge deployment; Liquid reports a 256K context for its 1.7B text model. Core capabilities include: Open-weight small models, Multimodal model family, On-device deployment.
Common scenarios include: Evaluate mobile inference, Build robotics model uses.
