Deploy local agents everywhere with LFM2.5-2.6B
LFM2.5-2.6B is a lightweight model designed for on-device agent deployment, offering efficient inference on various hardware platforms. It excels in instruction-following and tool-use benchmarks, outperforming smaller models while maintaining strong performance in knowledge and math tasks. The model achieves high inference speeds, such as 220 tokens per second on an Apple M5 Max and 113 tokens per second on an AMD Ryzen CPU, with low memory usage, making it suitable for mobile and portable devices.
More from Huggingface
LFM2.5-VL-3B for Better and Faster Vision Capabilities for the Edge
Thinking of ACE? We Can Do It with Fewer Tokens
Meta is back with Muse Glimmer: local, agentic, multimodal, and open source
