Liquid AI has released d1.2 and d1.2-ultra, decision models that skip token generation entirely: given a state plus typed questions, they return calibrated probabilities for every option in a single call, adding five new answer types including value reading, span finding, box drawing, range estimation and multi-select. d1.2-ultra scores 75 on the Artificial Analysis Decision Index v2.1, two points ahead of rivals at 73, with a 0.8-point calibration error versus 3.7 for Perplexity Decider v1.3. The trade-off: d1.2 scores just 69, trailing eight models, and parameters, latency and weights remain undisclosed.
Modelsslug: d1-2
d1.2
Liquid AI's lighter decision model adding five answer types including value reading, span finding and box drawing.
Curated Intelligence
Marktechpost (@Marktechpost)T2·Models
78 pts
Artificial Analysis (@ArtificialAnlys)T1·Models
62 pts
Liquid AI refreshed its flagship decision-model line with d1.2-ultra and d1.2. d1.2-ultra tops the AA-Paperclip private benchmark within Artificial Analysis Decision Index v2.1 at 76 points, three ahead of runner-up Drex Max v1.1, and both models sit on the capability-versus-cost Pareto frontier.