# LFM2.5: what I've logged

> LFM2.5 comes up in 2 posts of my AI log, from January 2026 to October 2026, and each item below links to the primary source. The sources come from Liquid AI, filed under open weights, agents and coding and image, video, world.

## Items

1. **LFM2.5-24B-A2B: Built for Fast On-Device Agents** (Liquid AI, 13 Oct 2026): An open-weight 24B reasoning model with 2B active parameters that scores 27.69 on AutomationBench against 8.72 for Qwen3.6-35B-A3B and runs at 184 tokens/s on an Apple M5 Max. [Primary source](https://www.liquid.ai/blog/lfm2-5-24b-a2b) · [Post](https://heyaliasif.com/ai/2026-10-15-liquid-lfm25-tokenspeed-llamacpp#item-1)
2. **LFM2.5** (Liquid AI, 5 Jan 2026): A new generation of small on-device models from Liquid AI, trained for running on phones and laptops. [Primary source](https://www.liquid.ai/blog/introducing-lfm2-5-the-next-generation-of-on-device-ai) · [Post](https://heyaliasif.com/ai/2026-01-05-week-in-ai#item-1)

## Posts

- 2026-10-15 [Liquid's 24B on-device agent model, TokenSpeed, llama.cpp updates](https://heyaliasif.com/ai/2026-10-15-liquid-lfm25-tokenspeed-llamacpp)
- 2026-01-05 [CES week: Nvidia's open driving model, Atlas for factories, LTX-2](https://heyaliasif.com/ai/2026-01-05-week-in-ai)

Canonical: https://heyaliasif.com/ai/topics/lfm2-5