This issue continues with gpt-4o fine tuning models + image input.
Chat Completions (1/1):
Responses API (@ 4/4)