Liquid AI releases LFM2.5-24B-A2B, a 24B-parameter model built to run on your laptop

Liquid AI has released LFM2.5-24B-A2B, an open-weight model with 24 billion parameters that the company says can hang with larger competitors. The twist is that it only uses a fraction of that size at any given moment. The model went live on October 13, 2026, and is available on Hugging Face. It is the biggest entry in the LFM2.5 family. It also caps the series, which Liquid AI has spent the year aiming at one goal: making capable AI run locally on consumer hardware instead of in a distant data center. Big model, small footprint The headline spec is a split personality. LFM2.5-24B-A2B holds 24 billion parameters in total but activates only 2 billion per token. That design is what makes the speed numbers possible. Liquid AI says the model reaches up to 184 tokens per second on Apple’s M5 Max chip. The company also claims the model runs up to 1.7 times faster than comparable mixture-of-experts models on edge devices. The context window sits at 131,000 tokens. The vocabulary covers 128,000 entries, and the model supports 16 languages. How it stacks up against bigger rivals According to Liquid AI, LFM2.5-24B-A2B leads or ties on 5 out of 8 agentic and tool-use benchmarks when measured ...