LiquidAI/LFM2-1.2B
LFM2-1.2B by Liquid AI is a 1.2 billion parameter hybrid language model with a 32,768 token context length, optimized for edge AI and on-device deployment. It features a novel architecture with multiplicative gates and short convolutions, achieving faster training and inference speeds compared to previous generations and similarly sized models. This model excels in knowledge, mathematics, instruction following, and multilingual capabilities, making it suitable for fine-tuning on narrow use cases like agentic tasks, data extraction, RAG, creative writing, and multi-turn conversations.
Loading preview...
LFM2-1.2B: A Hybrid Model for Edge AI
LFM2-1.2B is a 1.2 billion parameter hybrid language model developed by Liquid AI, specifically engineered for efficient edge AI and on-device deployment. It introduces a new architecture combining multiplicative gates and short convolutions, leading to significant improvements in speed and memory efficiency. The model boasts 3x faster training and 2x faster decode/prefill speeds on CPU compared to Qwen3, while maintaining a substantial 32,768 token context length.
Key Capabilities
- Optimized Performance: Outperforms similarly-sized models across various benchmarks including knowledge, mathematics, instruction following, and multilingual tasks.
- Efficient Deployment: Designed to run efficiently on CPU, GPU, and NPU hardware, enabling flexible deployment on smartphones, laptops, and vehicles.
- Multilingual Support: Supports English, Arabic, Chinese, French, German, Japanese, Korean, and Spanish.
- Tool Use: Features a structured tool-use mechanism, allowing for function definition, calling, execution, and interpretation within conversations.
Recommended Use Cases
LFM2-1.2B is particularly suited for fine-tuning on narrow use cases to maximize performance. It is recommended for:
- Agentic tasks
- Data extraction
- Retrieval Augmented Generation (RAG)
- Creative writing
- Multi-turn conversations
It is not recommended for knowledge-intensive tasks or those requiring extensive programming skills without further fine-tuning.