Build specialized multimodal models with LFM3
What changed
LFM3 is our new generation of Liquid Foundation Models (LFMs), built to be the starting point for specialized intelligence. LFM3 brings four main changes compared to our previous LFM2 family: - More capability per parameter: a new tri-hybrid architecture and training recipe allow LFM3-1.7B to outperform LFM2.5-2.6B while using 35% fewer parameters. Compared with LFM2.5-Audio-1.5B, it uses a new in-house audio encoder and speech codec, speaks with higher quality (UTMOSv2 3.83 against 3.22), and adds audio understanding, general-purpose audio reasoning, and tool calling.
Why it matters
A concrete addition to Practical AI: Tools, Models & Frameworks: it changes what's available to builders today rather than being general commentary.
How it compares
Related prior coverage to compare against:
- Ollama's new engine for multimodal models
- Multimodal open d1 decision models for the edge
- Multimodal Embedding & Reranker Models with Sentence Transformers
Sources
- Build specialized multimodal models with LFM3 (huggingface-blog)primary