Microsoft launches Microsoft-Decision-1, a fast decision‑scoring model trained on Qwen3.5‑9B, and plans to rebase it on MAI, OpenAI and other models
microsoft openai qwen
| Source: Techmeme | Original article
Microsoft has introduced Microsoft-Decision-1, a fast decision‑scoring model built on Qwen3.5‑9B, and plans to rebase it on MAI, OpenAI and other models.
Microsoft has rolled out a new AI model called **Microsoft‑Decision‑1**, a 9‑billion‑parameter system built by post‑training Alibaba’s Qwen3.5‑9B. Unlike conventional large language models that generate free‑form text, Decision‑1 is designed for “single‑pass decision scoring,” returning a calibrated probability for each fixed answer option. The model is already accessible through Microsoft Foundry and OpenRouter, with a usage price of $0.042 per million input tokens and no charge for output tokens.
The launch highlights a fast‑growing niche in artificial intelligence: decision‑oriented models that prioritize latency and deterministic scoring over generative flexibility. Microsoft positions Decision‑1 as a tool for automated classification, routing, and workflow governance—tasks such as triaging support tickets, flagging risky transactions, or selecting among a set of predefined replies. Early performance claims cite “top performance in latency and quality on structured decision tasks,” suggesting it can outpace both generic LLMs and existing decision‑specific solutions.
Microsoft also announced plans to “rebase” Decision‑1 on additional foundations, including models from its own Microsoft AI (MAI) portfolio, OpenAI, and other providers. This move could broaden the model’s capabilities, improve calibration, and align it with the broader ecosystem of large‑scale language models that Microsoft is integrating across its cloud services.
What to watch next is how quickly developers adopt Decision‑1 for enterprise workflows and whether the upcoming rebasing delivers measurable gains. The model’s pricing and availability on Foundry may spur competition in the decision‑scoring space, prompting rivals to release comparable offerings. Observers will also monitor how the model fits into Microsoft’s broader AI strategy, especially in light of recent leadership comments about model security and the need for “emergency brakes” on AI systems.
Sources
Back to AIPULSEN