We're releasing two flavors of these open models 我看下面有知友回答说GPT-5的生图比GPT-4o变差了，但是据我所知，GPT-5并没有直接的生图能力，生图还是接的基于GPT-4o的gpt-image-1，所以大家的判断还是有主观因素在的。 而且在 GPT-5 刚上线时就直接强制停用 GPT-4o，这更会让部分用户产生抵触情绪。 We have also released a dataset for researchers to study their behaviors. The easiest way to get started with benchmarking is by checking out the template repo, which contains detailed instructions and an agent template.
WATCH
