Liquid AI open-sources Open d1 decision models with zero output tokens
Liquid AI released the open-weight Open d1 family of decision models: d1-3B (3.12B parameters, reads text and images) and d1-omni-600M (587M, an early research version that reads text with either images or audio). Rather than generating text, the models read a state once and return typed, calibrated probabilities for named questions, with single-question latency quoted as low as 8 ms, aimed at routing, moderation, intent classification, reranking, LLM-as-a-judge scoring and agent guardrails. Liquid AI says d1-3B scores 48.57 on Decision Index v0.2.1 using its own official scorer (not a leaderboard submission), above all sub-10B models; both checkpoints are on Hugging Face under the LFM Open License v1.0, free for commercial use below $10M annual revenue.
Framing the output as calibrated probabilities instead of text could make real-time classification, routing and guardrails on edge devices cheaper than running a small chat language model, if the latency and calibration claims hold up.