Microsoft's new decision model charges nothing for output and $0.042 per million input tokens

Microsoft released Microsoft-Decision-1 on October 9, a model built only to pick among fixed options for tasks like classification, routing and verification. Instead of generating text, it returns a calibrated probability for each option in a single pass, covering yes/no, multiple choice, ratings and rubric-based grading of AI answers and agent actions. It is post-trained from Qwen3.5-9B, and Microsoft plans to rebase it on MAI and OpenAI models. It is available now in Microsoft Foundry and on OpenRouter at $0.042 per million input tokens, with output free.

Microsoft says it topped a 36-benchmark comparison of nearly 150,000 questions and ran 2.5 times faster than the runner-up, H2O-Lightning-4B, and 35 times faster than GPT-6 Sol. Xbox Research used it to sort more than 10,000 reviews into themes and found it competitive with GPT-6 Sol at 14 times the speed and a two-hundredth of the cost. The same day, Nace.AI released open weights for Drex 1.5, a decision model under 10 billion parameters that ranks first on the public Decision Index 0.2.1 at 58.28, ahead of Jev 1.13.0 at 57.91.