Microsoft Launches Decision-1, an AI Model Built for Fast Business Decisions

Microsoft introduced Microsoft-Decision-1 on Friday, October 9, 2026, describing it as a model built specifically for fast, structured decision-making rather than open-ended text generation, according to the company's official announcement on its Command Line site.

CEO Satya Nadella announced the launch on X, writing that the model "delivers top performance on structured decision tasks, outperforming both LLMs and other decision models in latency and quality," and said Microsoft was already testing it internally for tasks ranging from incident response to scientific discovery.

What Decision-1 Actually Does

Unlike large language models that generate free-form text, Microsoft-Decision-1 takes a fixed set of answer options — such as yes/no choices, multiple-choice items, rating scales or rubric-based grading criteria — and returns a calibrated probability score for each option through a single structured API call, according to Microsoft's product page.

Microsoft says the model was built by post-training Qwen3.5-9B, a 9-billion-parameter base model, for single-pass decision scoring. The company said it plans to rebase future versions on other models, including its own Microsoft AI systems and OpenAI models.

Benchmark Performance

Microsoft reports that Decision-1 achieved the highest accuracy across 36 benchmarks spanning nearly 150,000 questions that were kept blind from the model's training data. On latency, the company says the model was the fastest tested, running roughly 35 times faster than OpenAI's GPT-6 Sol and 2.5 times faster than the runner-up, H2O-Lightning-4B v1.1, at median (P50) response time, per Microsoft's announcement and reporting by TestingCatalog.

Microsoft also says the model changes its decision in only about 1.3% of cases when inputs are slightly reworded, a measure of consistency the company calls robustness.

Early Internal Use Cases

According to Microsoft's announcement, the Xbox Research team used Decision-1 to categorize more than 10,000 pieces of game feedback, finding output quality comparable to GPT-6 Sol while running more than 14 times faster and about 200 times cheaper. Microsoft listed 15 target applications overall, including agent controls, model routing, data labeling, incident-response routing, content classification and scientific-discovery optimization.

Availability and Pricing

Decision-1 is available now through Microsoft Foundry and is listed as coming soon on the third-party model marketplace OpenRouter, Microsoft said. Input tokens are priced at $0.042 per million tokens, while output tokens are offered free of charge.

Seeking Alpha reported that Nadella's announcement, made via a post on X, framed Decision-1 as part of Microsoft's broader push to embed AI decision-making directly into enterprise software workflows rather than relying solely on general-purpose chat-style models.