Reported News

Liquid AI releases LFM3, six open-weight 2B-class models for on-device text, vision and audio

Liquid AI released LFM3 on October 27, 2026: six open-weight models of about 2B parameters branched from one LFM3-1.7B-Base backbone pre-trained on 22T tokens. Liquid AI says it found "LFM3-1.7B to outperform our previous LFM2.5-2.6B while using 35% fewer parameters".

Share on X
Changes reported
4
listed below, each from a source we read
We tested
0
we did not run it ourselves
Gaps we flag
2
things we could not verify
Sources
3
3 primary, linked below

What changed

Changes the sources report, with the evidence status of each
#What the source saysStatus
01The family covers text (LFM3-1.7B), vision (LFM3-VL-2B), audio (LFM3-Audio-2B) and omni (LFM3-Omni-2B), plus two decision models specialized from the vision and omni models.Reported
02All six branch from LFM3-1.7B-Base, pre-trained on 22T tokens on a new hybrid architecture.Reported
03The text model has a 256K-token context window and fits in about 1 GB at 4-bit precision.Reported
04Weights are on Hugging Face under the LFM Open License v1.0, with GGUF, MLX, ONNX, FP8 and NVFP4 builds.Reported

reported by the sourcemeasured or run by usnot verified

Primary source: Liquid AI blog: LFM3. We are the fast, independent index; the vendor’s own page is the source of record for this release.

In the source’s words

“Built for the edge: the text model fits in about 1 GB at 4-bit precision.”

“LFM3-1.7B to outperform our previous LFM2.5-2.6B while using 35% fewer parameters”

What we have not verified

  • Performance comparisons are Liquid AI's own; we did not benchmark LFM3.
  • We did not run the models on a laptop or phone.

This brief is Reported: written from the sources below, not from running the release ourselves.

Sources

Spotted an error? See our corrections policy or email the address on the contact page. Markdown version for agents: /news/liquid-ai-lfm3-on-device-model-family.md

Want these briefs in your Google Top Stories? Add Agentic Weekly as a preferred source on Google.

The Weekly Verdict · free

5 new-model use cases. Every Monday.

A free email every Monday: the week's few big AI releases, then 5 short use cases for the new models, each with a prompt you can paste.

Free. Double opt-in. One click to unsubscribe.