M-JudgeBench: Benchmarking Multimodal Large Language Models as Judges

M-JudgeBench: Benchmarking Multimodal Large Language Models as Judges — Arxiv

OpenAI pauses development of Astra AI model over critical cybersecurity capability concerns