Verified, then logged.
Sun 22 Nov 2026 · 00:34 KST / 15:34 UTC한국어 | ENSearch /

Breaking02 / IT

Ollama adds Liquid AI's LFM3 models, on Apple hardware first

The release notes do not name a licence or give a date for Linux and Windows support.

On a desk at dusk, a silver laptop glows with blurred terminal output while a black desktop PC and monitor sit powered off behind it. AI-generated image

Ollama added Liquid AI's LFM3 models in v0.40.12 on 27 Oct 2026, and the release notes name no licence for the family, which runs on Apple hardware for now.

According to the release notes, lfm3 is a 1.7B text model with a 256K-token context window, the length of input a model can process at once. The lfm3-vl model adds image understanding.

The lfm3-audio model handles speech recognition, translation and text-to-speech in 18 languages, the release notes state. The lfm3-omni model handles text, images and audio in one model.

Ollama describes LFM3 as a family of small models for local agents, tool calling and structured output. The family is built on Liquid's new tri-hybrid architecture. The run command is ollama run lfm3.

Liquid AI's d1 models now come in a 2B size built on LFM3, according to the release notes. The d1 model makes decisions about text and images, and d1-omni adds audio.

The list of changes in the release includes faster prompt processing on Apple silicon for Gemma 4, Flint and Laguna models. The ollama launch command now works with OpenCode v2.

Ollama said Linux and Windows support for lfm3 and the 2B d1 is coming soon. The release notes state neither a date for that support nor a minimum VRAM.