Liquid AI launches d1.2 decision models with text, image and audio support
Liquid AI has released d1.2 and d1.2-ultra, hosted decision models that process text, images and audio. Instead of generating prose, they return probabilities for predefined answers and typed results, producing zero output tokens. The models support up to 65,536 input tokens per question and 128 questions in one call. Their parameters are undisclosed, and both are currently available through the hosted Liquid API only.
The update adds five response types: extracting a value with its source text, finding every matching span, locating objects with bounding boxes, estimating a number with a range, and selecting multiple applicable options. Audio is provided as 16-bit PCM WAV and is charged at 10 tokens per second. On held-out ESC-50 and FLEURS recordings, the models answered 95% to 97% of questions correctly.
On the Artificial Analysis Decision Index v2.1, d1.2-ultra scored 75, the highest score among the decision models listed in the article; d1.2 scored 69. In Liquid AI’s own evaluation of 13 benchmarks and 22,505 questions, calibration error was 0.8 points for ultra and 1.2 for d1.2. Weights are planned but have no announced release date. Earlier d1 models are deprecated and lack audio and the new tasks.
Written by Sakhanda Wire AI from reporting by MarkTechPost. Facts, quotes and full context are in the original article below.