Microsoft Launches 9B Qwen Decision Model

Microsoft unveiled a 9-billion-parameter AI on Oct 9 2026 that outright refuses to write paragraphs. Named Microsoft-Decision-1, it scores a fixed list of answer options in a single forward pass. It then returns a calibrated probability for each option. The model is built on Alibaba's open-weight Qwen3.5-9B and was dropped on Microsoft Foundry immediately, with OpenRouter access following shortly. Microsoft pitches it for routing, classification, prioritization, verification and workflow control in high-volume, low-latency tasks. In internal tests it hit 83.5 percent accuracy across 36 benchmarks and about 150 000 held-out questions. Microsoft says Decision-1 runs 4.5 times faster than Quyet-1.0-Large, 35 times faster than GPT-6 Sol, and about 85 ms latency. "We'll soon rebase it on other models, including Microsoft AI and OpenAI," the firm said.