Liquid AI d1.2 and d1.2-ultra decision models available on AI Gateway with audio and bounding box support

·2026.11.19 09:00

Key point

The models support Boolean, Choice, and Score questions, returning structured answers with audio input and image bounding box capabilities.

Details

  • Liquid AI's d1.2 and d1.2-ultra are now available on AI Gateway, supporting audio input and bounding box detection.
  • These decision models return structured answers (yes/no probabilities, category selection, or rubric ratings) instead of generating text, allowing multiple questions to share one input.
  • d1.2 is the faster, lower-cost option, while d1.2-ultra is Liquid AI's most accurate decision model.
  • Usage requires setting the AI_GATEWAY_API_KEY and using model names liquid/d1.2 or liquid/d1.2-ultra.
  • Audio classification requires ai 7.0.136 or later, passing audio as a file part in state.
  • Bounding box coordinates are returned as fractions of image height and width via the Decisions API.
  • Decision requests appear in AI Gateway logs and count toward budgets.

This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.

Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.