Liquid AI has released two omnimodal decision models, d1.2 and d1.2-ultra, designed to process text, audio, and image inputs for data extraction and numerical estimation. The top-tier d1.2-ultra model claimed the No. 1 spot on the Artificial Analysis Decision Index v2.1 with a score of 75, while the standard d1.2 model ranked sixth on the same index with a score of 69.
The flagship d1.2-ultra also topped the private AA-Paperclip benchmark with a score of 76, a 4 point improvement over its predecessor, d1-ultra. Both models are available via Liquid AI’s API and feature a 64K token context window, with input pricing starting at $0.02 per 1M tokens for the d1.2 version.
