
Why LFM3?
Built on a new tri-hybrid architecture, LFM3 sets a new standard in quality, speed, and memory efficiency.More Capability per Parameter — New tri-hybrid architecture is more expressive while keeping LFM2’s speed
State-of-the-art Quality — Outperforms similar-sized models on benchmarks
Memory Efficient — Optimized for resource-constrained environments
Deploy Anywhere — Compatible with major inference frameworks and platforms
Get Started
Explore Models
Browse text, vision, audio, omni, and decision models with their formats and capabilities
Inference Guides
Learn how to run models for different use cases and platforms
Fine-tuning
Customize models for your specific requirements
Examples
End-to-end examples for mobile, laptop, and web
Model Families
Text Models
Chat, tool calling, structured output, and classification
Vision Models
Image understanding with LFM backbones and custom encoders
Audio Models
Interleaved audio/text models for TTS, ASR, and voice chat
Omni Models
Text, image, video, and audio input with text or speech output in one model
Decision Models
Classification, routing, scoring, and extraction with zero generated tokens: d1.2-ultra and d1.2 via the Liquid API, open-weight d1-2B, d1-omni-2B, and d1-omni-600M
Liquid Nanos
Task-specific models for extraction, retrieval for RAG (embeddings and reranking), summarization, and translation