Model1h ago
Hugging Face Releases LFM2.5-2.6B Local Model
The 2.6 billion parameter LFM2.5 model enables high-performance local agent deployment on consumer hardware. It outperforms larger models in tool-calling and reasoning tasks while maintaining a small memory footprint. Hugging Face optimized this architecture for low-latency edge execution. Developers can now run complex agentic workflows locally without relying on expensive cloud API calls.