Liquid AI announces six LFM3 models aimed at on-device use

The six-model LFM3 family spans text, vision-language, audio, combined modalities and decision tasks, according to Liquid AI. The company pitches single-GPU fine-tuning and on-device use, alongside features including a 256k context window for its text model. Separately, @atomic_chat_hq claims its quantized versions of that model fit on an 8GB Mac. The hardware fit claim comes from the poster, not an independent test.