Microsoft ships Decision-1, a model that scores choices

What happened: Microsoft introduced Microsoft-Decision-1, a decision model for routing, classification, prioritization, verification and workflow control. Instead of generating text, it returns a calibrated probability for each fixed answer option, supporting yes or no, multiple choice, rating and rubric grading through a structured API call. It is available now in Microsoft Foundry and on OpenRouter, priced at $0.042 per million input tokens with output tokens free.

Context: Microsoft post-trained the model from Alibaba’s Qwen3.5-9B. It says Decision-1 led a 36 benchmark comparison across nearly 150,000 questions and was the fastest measured, about 2.5 times quicker than the runner-up H2O-Lightning-4B and 35 times quicker than GPT-6 Sol. The company reports it flips a decision on 1.3% of perturbations on average when inputs are paraphrased or reordered, and says it will rebase the model on its own MAI models and on OpenAI.

Why it matters: Decision models are emerging as their own category, and cheap structured output is exactly what agents need between steps, to route, check and grade without paying a frontier model to write a paragraph. Microsoft’s own examples, labeling more than 10,000 pieces of XBOX feedback or replanning scientific experiments, show the target is the plumbing around models rather than the models themselves.

(Original source: Microsoft Command Line)