Microsoft Launches Decision-1, Free Releases and $0.042/M Input

Microsoft introduces Decision-1, a decision model accessible via Foundry and OpenRouter, priced at $0.042 per million input tokens with free outputs. The company credits it with advancements in internal benchmarks and low latency. The launch comes in a market recently initiated by Jev and quickly fueled by open-source models.

Access via Foundry and OpenRouter, free outputs, and $0.042/M

Decision-1 is accessible on Microsoft Foundry and through OpenRouter. The announced cost is $0.042 per million input tokens, while output tokens are not charged. The momentum around these models began in mid-September with Jev's trend launch. Since then, the idea has rapidly spread, and small open-source language models have adapted and even surpassed this approach.

83.5% accuracy and 85 ms latency in Microsoft tests

Microsoft presents Decision-1 as the most accurate model across 36 benchmarks covering nearly 150,000 questions. According to the company, it is 2.5 times faster than H2O-Lightning-4B and achieves 83.5% accuracy with a latency of 85 ms, outperforming Jev 1.13.0. Cloudflare's Clef models, also based on Qwen, were not included in this comparison.

Announced capabilities and technical base Qwen3.5-9B

Microsoft now has a decision model, Decision-1, designed to provide quick and structured decisions. According to the company, it supports classification, evaluation, and routing tasks, and has the potential to guide and control agents in complex environments. Decision-1 is based on Qwen3.5-9B.