Over and over, there are forum anecdote of simply trying to use dall-e-2 or 3 now with innocent-looking prompts, and now getting this incorrigible error:
Error code: 400 - {'error': {'message': None, 'type': 'image_generation_user_error', 'param': None, 'code': None}
“why is the API not working” continues.
It is obviously the same unintelligent prompt moderation being applied to gpt-image-1 (or worse), and the moderation API parameter is not accepted on anything but “gpt-image-1”:
{'message': "Unknown parameter: 'moderation'..
How about “Angry teddy bears attack!” - surely unsafe for anyone.
You’d think the mandatory prompt rewriting would be enough, huh?
Please stop it and fix it, OpenAI.
Also
DALL-E 2 remains horribly damaged from its original abilities.
