Microsoft Debuts Decision Model Built on Alibaba's Qwen

Microsoft has entered the race for "decision models"—AI designed for structured outputs rather than open-ended text—with the launch of Microsoft-Decision-1, as first reported by The Register.

While the company is positioning the tool as a high-performance, low-cost alternative for business applications, the model's current foundation is external. The Register reports that Microsoft-Decision-1 is based on Qwen3.5-9B, a model developed by Alibaba Cloud, the cloud computing division of Chinese tech giant Alibaba.

Achint Srivastava, VP of software engineering in the Office of the CTO at Microsoft, stated in a Friday blog post that decision models are a critical new category capable of delivering outputs that software can act upon immediately. Microsoft claims the model is 2.5x faster than H2O-Lightning-4B and 2.8x faster than TypeSafe AI's Jev in latency tests, while maintaining 83.5 percent accuracy across 36 benchmarks. The company also notes the model is over 20x cheaper than OpenAI's GPT-6 Sol for text classification, with input tokens priced at $0.042 per million and free output tokens.

Despite the reliance on Alibaba's architecture, Microsoft told The Register it intends to rebase Decision-1 on its own internal models and those from OpenAI in the future, though the company did not disclose the reasons for the initial reliance on Qwen. Microsoft-Decision-1 is currently available via Microsoft Foundry and will soon be offered through OpenRouter.