TL;DR: Decision 3.0, a new family of open-source multimodal decision models, has been released, offering state-of-the-art performance across various sizes and input types.
Summary: The vLLM Semantic Router team has released Decision 3.0 (d3), an Apache-2.0 licensed family of multimodal decision models ranging from 0.6B to 27B parameters. Built on Qwen3.5 and Qwen3.8-27B, these models process text, images (up to 1.6 MP), and video (sampled at 2 fps, up to 32 frames) to provide choice, yes/no, or score decisions with probabilities, without text generation. The 27B model achieved the #1 rank on both Jev Decision Index boards for text and vision, with strong zero-shot video performance.
Why it matters: This release provides indie developers and AI entrepreneurs with powerful, open-source tools for building applications requiring multimodal decision-making. Experiment with the different model sizes to find the optimal balance of performance and resource usage for your specific AI projects, especially those involving complex visual and video analysis.
Source: reddit