Microsoft Decision 1

Microsoft-Decision-1 is a decision model from Microsoft, built by post-training Qwen3.5-9B for fast, single-pass decision scoring. Given a fixed set of answer options, it returns a calibrated probability score for each option. It supports yes/no, multiple-choice and rating options, as well as rubric-based grading of AI responses and agent actions, through a structured API call.

Microsoft positions it for routing, classification, prioritization, verification and workflow control inside applications, agents and workflows. According to Microsoft's own benchmarks (36 benchmarks, nearly 150,000 questions kept blind from training), it had the highest accuracy and lowest latency among compared models, about 2.5x faster than H2O-Lightning-4B v1.1 and about 35x faster than GPT-6 Sol. It reportedly flips its decision on 1.3% of input perturbations. These figures are vendor-reported and not independently verified here.

It is available through Microsoft Foundry and OpenRouter. Microsoft says it plans to rebase the model on other models, including Microsoft AI (MAI) and OpenAI models. Internal uses cited include labeling game feedback for Xbox Research, quality control of Copilot responses, incident-response knowledge retrieval and adaptive replanning in Microsoft Discovery. It outputs structured scores rather than free-form text.