#Thinking mode gets confused when asked to generate both text and an image

1 messages · Page 1 of 1 (latest)

spring vaultBOT
#

Reported by @stone siren

Bug Report: Thinking mode gets confused when asked to generate both text and an image
`Steps to Reproduce`
  1. Select GPT-5.2 Thinking
  2. Ask the AI to write a story and generate an image to go with it, in a single message
`Expected Result`

The AI should have no problem fulfilling the request. It should also be able to do the story and the image in the correct order, if a specific order was requested.

`Actual Result`

The AI may or may not fulfill the request correctly. Either way, it will often spend a lot of time thinking about how to fulfill the request. The AI seems to think it can’t generate an image without leaving the message empty even though it’s perfectly capable (and sometimes even falsely thinks I asked it to leave the final message blank.) The image will often be generated first even if the request was to generate the story first.

`Environment`

iOS 26.0.1 / macOS Safari

#
Additional Information

Please provide relevant details to help resolve the issue, such as:

  • ChatGPT Shared Link (if applicable).
  • Screenshots or videos demonstrating the problem.

-# ➜ Need to contact support? Visit the OpenAI Help Center.