Microsoft Tests Low-Cost AI Decision Model in Internal Teams

Image: Collected

Microsoft says its new Decision-1 AI model has been tested by teams across the company, where it was used to sort player feedback, evaluate Copilot responses and support scientific research.

In one internal trial, Xbox Research used Decision-1 to classify more than 10,000 pieces of player feedback into themes. Microsoft said the model delivered quality comparable to GPT-6 Sol while running more than 14 times faster and at one-two-hundredth the cost.

Microsoft’s Copilot team found Decision-1 competitive with GPT-5.6 Luna for judging response quality, while operating 100 times faster, the company said. In another trial, Microsoft Discovery reported that Decision-1’s scoring was 46 times more consistent than an approach based on a large language model.

Microsoft is positioning Decision-1 as a low-cost tool for the many small judgments AI agents make while completing tasks. The company’s reported results come from internal trials and have not been independently verified.