Microsoft has introduced Microsoft-Decision-1, a specialized artificial intelligence model designed to make structured decisions faster and at a lower cost. The company claims the model delivers median-latency performance approximately 35 times faster than GPT-6 Sol, positioning it for AI agents, workflow automation and business applications.
Unlike general-purpose AI models that generate text and answer complex questions, Decision-1 focuses on selecting predefined options and assigning probability scores to possible outcomes. The aim is to handle routine decisions without repeatedly relying on larger, more expensive language models.
Built on the Qwen3.5-9B model, Decision-1 has undergone specialized post-training to improve decision scoring. According to Microsoft's published findings, it is also 4.5 times faster than Quyet-1.0-Large in median-latency performance.
Microsoft said Decision-1 ranked first in accuracy across 36 benchmark tests covering nearly 150,000 questions. However, these results have been reported by the company and have not been independently verified.
The model is priced at USD 0.042 per million input tokens, while output tokens are free. The pricing is intended to make repetitive, high-volume decision-making tasks more economical for businesses and developers.
Microsoft's Xbox Research team tested Decision-1 on more than 10,000 pieces of game feedback and reviews. The company said the model delivered quality comparable to GPT-6 Sol while operating more than 14 times faster and at approximately one-two-hundredth of the cost.
Decision-1 is good for yes/no decisions, multiple choice, rating and rubric assessment. Uses may include content classification, job application routing, job prioritization, quality control, validation of outcome, and safety checks.
For AI agents, Decision-1 will determine whether the job should be done, repeat the action, escalate the issue, or require human intervention. These capabilities could reduce processing times and operating costs in applications that handle large volumes of decisions.
Also Read: Oracle vs Microsoft: Comparing Two AI-Driven Cloud Stocks
Microsoft said it tested Decision-1 for harmful-content identification, jailbreak attempts and prompt injection. The company also evaluated its consistency under different input variations, reporting an average decision-change probability of 1.3% across eight perturbation methods.
The findings are based on Microsoft's evaluations and require independent validation. Decision-1 is available to developers through Microsoft Foundry, while access through OpenRouter is planned.
The launch reflects Microsoft's effort to separate routine decision-making from more demanding reasoning and text-generation tasks. Independent assessments will be needed to establish whether the reported speed, accuracy and cost advantages hold up across real-world applications.