📢 Really excited to host the Data Curation for Vision Language Reasoning Challenge (DCVLR) @ NeurIPS 2025 and to include VMCBench as one of the evaluation sets! We’re looking forward to seeing the top solutions (with prize!)—huge thanks to @benjaminfeuer.bsky.social and the team for their hard work!

🔍 Vision language models are getting better - but how do we evaluate them reliably? Introducing AutoConverter: transforming open-ended VQA into challenging multiple-choice questions! yuhui-zh15.github.io/AutoConverte...