Microsoft launches Microsoft-Decision-1, fast decision-scoring AI model for agent workflows
What happened
Microsoft announced Microsoft-Decision-1 on October 9, 2026 — a new 'decision model' that scores a fixed set of options with calibrated probabilities instead of generating text, designed for routing, classification, prioritization, verification and agent-control tasks. The model is available now in Microsoft Foundry (and due on OpenRouter soon), built via post-training on Alibaba's Qwen3.5-9B, and priced at $0.042 per million input tokens with output tokens free. Microsoft claims the highest accuracy across 36 benchmarks covering nearly 150,000 questions, 4.5x faster than runner-up Quyet-1.0-Large and 35x faster than GPT-6 Sol at P50 latency. Internal testing: Xbox Research categorized 10,000+ gaming feedback entries at comparable quality, 14x speed and 200x lower cost versus GPT-6 Sol; the Copilot team found it competitive with GPT-5.6 Luna on AI-response evaluation. MSFT gained about 2.3% on Friday and ended the week 3.4% higher.
Why it matters & what’s next are part of Pro, along with the audio summary.
Start your free 14-day trialWho’s affected
- MSFT — directly exposed: new Foundry-monetizable product targeting enterprise agent workflows; reinforces AI commercialization narrative
- NVDA — second-order: decision-model inference demand at Azure scale supports GPU consumption
- BABA/Alibaba — Qwen3.5 is the post-training base, a validation of its open models