Microsoft's Nadella Unveils Microsoft-Decision 1, A New AI Model For Fast Decisions

• Evaluated across 36 blind benchmarks, Microsoft-Decision-1 recorded 4.5 times faster execution than its closest competitor and ran roughly 35 times faster than traditional large language models like GPT-6 Sol. 
• The system focuses strictly on generating calibrated probability outputs for multiple-choice, rating, and pass-fail decisions rather than general text generation. 
• Available immediately via Microsoft Foundry (and coming soon to OpenRouter), the model costs $0.042 per million input tokens, with output tokens provided at no charge. 

Microsoft Corp.(MSFT) announced Friday the launch of a specialized artificial intelligence system designed to execute rapid automated decisions, marking a strategic pivot toward lower-cost, highly targeted AI tools for enterprise workflows.

The new software, titled Microsoft-Decision-1, is engineered to handle tasks such as sorting incoming queries, routing system requests, verifying data, and managing autonomous agents. 

Unlike conventional large language models built around conversational text, decision models are designed to evaluate defined options and output direct, structured metrics that software programs can consume without human intervention.

Achint Srivastava, Vice President of Software Engineering at Microsoft's Office of the CTO, said in a statement that decision-scoring systems represent an emerging category intended to address latency bottlenecks in software automation.

MSFT stock gained about 2.3% on Friday and ended the week 3.4% higher. 

Overcoming Speed And Accuracy

A key challenge in multi-step AI operations is accumulated delay. When multiple automated decisions execute sequentially, slight lags at each step can slow down entire applications.

To tackle this, Microsoft engineers post-trained a 9-billion-parameter base model (Qwen3.5-9B). In benchmark testing spanning nearly 150,000 blind evaluations, the system recorded the highest accuracy among evaluated models while operating at roughly 35 times the speed of flagship text models such as ChatGPT’s GPT-6 Sol.

The model also demonstrated stability against prompt manipulation. Internal evaluations showed the system changed its decision output in only 1.3% of cases when requests were paraphrased or when option orders were re-sorted.

Real-World Applications And Costing

Microsoft reported that several internal teams have already deployed the model to streamline operations. For instance, Xbox Research used the tool to categorize more than 10,000 player reviews into structured categories, achieving a 14x speed increase over previous setups.

The Microsoft Copilot team used the system for quality assurance checks on interactive agent responses, operating up to 100 times faster than prior configurations.

Microsoft-Decision-1 is accessible to developers through Microsoft Foundry, with plans to expand availability to OpenRouter. Input token processing costs $0.042 per million tokens, while output tokens are free. Microsoft said future iterations of the software will incorporate alternative model foundations, including OpenAI technology.

MSFT Stock: Retail View 

Retail sentiment on Stocktwits was ‘neutral’ with ‘normal’ message volumes. 

MSFT stock has gained about 10.3% year-to-date.