LiquidAI Releases LFM3 Series Multimodal Edge Models
Understand the story
On October 28, 2026, Liquid AI released the new-generation LFM3 series of open-weight models on Hugging Face. The series derives six 2B-class models from LFM3-1.7B-Base, using the same backbone to cover text, vision, audio, omni-modal, and decision-making tasks. Notably, LFM3 introduces its first omni-modal model and enables the vision model to support video input. The models are currently available on Hugging Face as open weights.
AI overview of reporting · Updated 1 h前
Reporting timeline
Explore different aspects of the story through its reporting.
- Hugging Face BlogSelectedBuild specialized multimodal models with LFM3
Liquid AI released the new LFM3 series of open-weight models on Hugging Face. Based on LFM3-1.7B-Base, it derives six 2B-class models covering text, vision, audio, omni-modal, and decision tasks, including the first omni model and vision models that support video input.
Story heat trend
Not enough continuous observations to show a trend yet.