Meituan LongCat team releases UNO-Bench, a one-stop benchmark for all-modal performance evaluation.

Meituan's LongCat team has launched UNO-Bench, a benchmark for evaluating multimodal large models, which accurately measures the performance of models on unimodal and multimodal tasks such as image, audio, video, and text. Built with high-quality, diverse data, the benchmark is the first to validate the "combinatorial law" of multimodal large models, revealing the complex relationship between unimodal and multimodal capabilities.