Microsoft Launches Next-Generation Decision Model: Over 2x Faster, Tops Accuracy Charts
1 day ago / Read about 0 minute
Author:小编   

Microsoft has released its next-generation decision model, Microsoft-Decision-1.1, which ranks 8th with 68 points on the relevant decision index leaderboard—a 6-point improvement over the previous generation. It offers more than 2x faster computational speed, lower costs, and added multimodal support. The model is priced at $0.042 per million input tokens, with output provided free of charge. Among 8 decision models across 102 benchmark tests, it achieves the highest average accuracy, and more than 4,000 customers have adopted it within just over five weeks of launch. During the same period, Microsoft also released multiple programming models for coding tasks. Microsoft achieved optimizations in latency, cost, and other dimensions through a custom inference engine, among other measures, and the model has been integrated into its own products such as Copilot. Microsoft is taking a differentiated approach by deeply integrating its models with its own products, enabling enterprises to connect directly to existing workflows with lower onboarding costs.