LiquidAI: LFM2.5-1.2B-Instruct (free)
🧠 AI Modelliquid
Compact 1.2B instruction model for fast on-device AI with 32K context.
LFM2.5-1.2B-Instruct is a compact instruction-tuned model optimized for on-device deployment. With only 1.2B parameters, it achieves competitive chat performance while maintaining low latency and memory footprint suitable for edge devices. It supports a 32,768 token context window and is fully text-based. The model is freely available on OpenRouter, providing a cost-effective solution for text generation tasks. It includes standard inference parameters like temperature, top_p, frequency penalty, and more, making it flexible for various applications. Its small size enables broad runtime support across platforms, from mobile CPUs to NPUs.
💡Highlights
- ├─1.2B parameters, 32K context
- ├─Free via OpenRouter
- └─Optimized for on-device inference
🎯For
- ├─Edge AI developers
- ├─Mobile app engineers
- └─Privacy-conscious users