Model behavior feedback

We’d love to hear examples where our models did not meet your expectations. Sharing submissions helps our research team improve our models.

Please note that by filling out this form, you agree to allow OpenAI to use this data to improve the model⁠(opens in a new window) via training and evaluations.