Liquid AI releases open-weight d1 decision models with vision and audio support
d1-3B and the experimental d1-omni-600M return structured decisions in a single forward pass instead of generating text; both are downloadable from Hugging Face.
2 min readRead story



