VideoAutoArena has been introduced as an automated and scalable benchmark designed for video understanding, aligning its rankings with human judgment to address complex video analysis tasks. The platform aims to go beyond traditional multi-choice questions in evaluating open-ended video analysis. Notably, the open-source model Aria has achieved the top ranking in VideoAutoArena, marking a significant milestone in the assessment of multimodal models for video analysis. This new benchmark is part of a broader trend in developing tools such as MMGenBench and VBench++, which focus on evaluating various aspects of large multimodal models, including text-to-image and text-to-video generation capabilities.