Liquid AI has released LFM2.5-2.6B, a 2.6-billion-parameter model built for on-device agentic use cases. The model goes through a training pipeline of supervised fine-tuning, teacher specialization, multi-domain distillation, and agentic reinforcement learning inside real agent environments. Liquid AI reports the model matches or beats models up to four times larger on instruction-following and tool-use benchmarks. It runs at 220 tokens per second on an Apple M5 Max and 113 tokens per second on an AMD Ryzen CPU while using under 2.5GB of memory, enabling deployment from laptops to phones.
