I was talking about reporting on testing and capabilities. Yes, open models provide a greater amount of information about the development of the model and how to run it yourself, but I am quite confident that literally no AI company, open or closed, conducts and reports so thoroughly on testing about the capabilities of their models.