Liquid AI has released two open-weight decision models, d1-3B and d1-omni-600M, designed for fast, structured decision-making on edge devices. Unlike generative models that produce tokens, these models answer in a single forward pass using Liquid Foundation Models (LFMs).

  • d1-3B scores 48.57 on the Decision Index 0.2.1, outperforming larger models like Decider 35B-A3B.
  • d1-omni-600M supports text with images or audio, surpassing Decider 2B with only a quarter of the parameters.
  • d1-3B achieves inference speeds under 50 ms on NVIDIA Jetson devices and under 10 ms on GPUs.

These models are available on Hugging Face for users needing low-latency multimodal decisions.