Liquid AI · Decision Model API

Liquid AI d1.2 and d1.2-ultra

Liquid AI has released d1.2 and d1.2-ultra, hosted decision models that return calibrated probabilities over fixed answer sets rather than generating text. Both accept text, image and audio input and add five new answer types: values, spans, boxes, ranged estimates and multi-label choices. d1.2-ultra scores 75 on the Artificial Analysis Decision Index v2.1, the top score among decision models on that index, and d1.2 scores 69; calibration error is 0.8 points for d1.2-ultra on Liquid AI's own 13-benchmark test. Access is through the Liquid API only, with no weights yet and parameter counts not disclosed; the page says Liquid AI will open weights "at some point".

Open the source ↗

Similar Language Models releases

More from Liquid AI

  • LiquidCodec

    Liquid AI · Streaming Audio Codec · Week 47, 2026

  • LFM3

    Liquid AI · Small multimodal LFMs · Week 45, 2026

  • LFM2.5-24B-A2B

    Liquid AI · Efficient Agentic MoE · Week 43, 2026

  • LFM2-24B-A2B

    Liquid AI · MoE Architecture · Week 9, 2026

  • LFM2.5-1.2B-Instruct

    Liquid AI · Small Language Models · Week 2, 2026

  • LFM2-2.6B-Exp

    Liquid AI · Small Language Models · Week 52, 2025

Published by Quanta AI Labs, an AI agency in India. Browse every week in AI News Weekly.