AI BriefingLiquid AI releases LFM3 family with 6 models across text, audio, vision, and omni modalities
·2026.10.28 02:23
Key point
The release includes LFM3-1.7B with 256k context and LFM3-VL-2B, which is claimed to be competitive with models up to 6x larger.
Details
- Liquid AI released LFM3, a new generation of Liquid Foundation Models comprising six instances across text, audio, vision, and omni modalities.
- LFM3-1.7B is a text model with 256k context, positioned as #1 on the Pareto frontier of on-device intelligence.
- LFM3-VL-2B is a 1.9B parameter vision-language model with video input for spatial understanding, claimed to be competitive with Gemma 4 12B, Gemini 3.5 Flash Lite, and GPT-5.6 Luna.
- LFM3-Audio-2B supports voice chat, tool calls, and reasoning in 18 languages, while LFM3-Omni-2B combines vision, audio, and language in a single model.
- d1-2B and d1-Omni-2B are decision models featuring typed outputs and calibrated probabilities in a single pass, with the Omni variant adding audio input.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Opposable runs 2B GUI-Owl 1.5 model on Android for on-device app control