Liquid AI adds audio and structured extraction to d1.2 decision models

Liquid AI adds audio and structured extraction to d1.2 decision models
By guthlabs.newsroom · November 16, 2026
Liquid AI has released d1.2 and d1.2-ultra, expanding its decision-model API beyond selecting an answer to extracting values, locating objects and interpreting audio. Both models are available through Liquid’s API now; OpenRouter and Vercel support is announced as coming later.
The new capabilities include reading text, numbers and dates; identifying matching passages; returning object boxes; estimating quantities with ranges; and selecting multiple applicable options. Audio can supply either the material being analyzed or the spoken question.
Liquid lists input-token prices of $0.02 per million for d1.2 and $0.042 for d1.2-ultra. The company says requests bill input tokens only, with images converted into token charges and additional questions adding their own text rather than rebilling the entire shared input.
Liquid reports that d1.2-ultra leads the Artificial Analysis Decision Index v2.1. Its claims for the five new tasks rely on internal evaluations, however, rather than a public benchmark for those capabilities.
The practical distinction from a general chatbot is structured output: a value, location, selection or confidence estimate that software can consume. Teams should still test those outputs on their own data before using confidence thresholds to automate consequential decisions.
Sources and citations
Revision history
Published .