Microsoft Unveils Decision-1: A Specialized AI Model for Rapid Decision-Making

Microsoft has ventured into a novel realm of artificial intelligence with the introduction of Microsoft-Decision-1. This unique AI model diverges from traditional large language models as it does not produce text. Instead, it evaluates predefined answer choices and provides a calibrated probability for each. This makes it ideally suited for swift decision-making in automated AI workflows.

The model was launched on October 8–9, 2026, and is now widely accessible on Microsoft Foundry. Support for OpenRouter is also on the horizon. The model, built on Alibaba’s Qwen3.5-9B base and further trained by Microsoft engineers, is tailored for tasks such as routing incoming queries, classifying tickets, verifying data, and managing autonomous AI agents.

In nearly 150,000 blind evaluations conducted during internal benchmark tests, it exhibited the highest accuracy among all models tested. It operates approximately 35 times faster than leading text models like OpenAI’s GPT-6 Sol and 4.5 times faster than rival decision models.

Microsoft’s competitive pricing strategy makes this model especially appealing. It is priced at a mere $0.042 per 1 million input tokens, with output tokens offered at no cost. This move indicates a wider industry shift towards lean, task-specific AI models that value speed and cost-efficiency over general-purpose text generation.

Source: NewsBytesApp — Microsoft introduces new AI model for making quick automated decisions