#OpenAI serves GPT-4o Mini even with GPT-5 Thinking selected

1 messages · Page 1 of 1 (latest)

digital patioBOT
#

Reported by @flat knot

Bug Report: OpenAI serves GPT-4o Mini even with GPT-5 Thinking selected (concrete proof in screenshot below)
`Steps to Reproduce`

(See screenshot in the replies below) The setup is in red, while observations are in yellow. The only way to reproduce this is to ask users who suspect they are intermittently encountering GPT-4o Mini (fast responses, no thinking, no tools) to repeatedly run a test. Choose a prompt which GPT-5 models would refuse but GPT-4o mini would comply (I'm not going to reveal mine since it is useful for holding companies accountable).

`Expected Result`

If it was GPT-5, it would have refused to reveal its system prompt. It would have reasoned, given that both GPT-5 Thinking + Extended Thinking were selected.

`Actual Result`

The model revealed its system prompt, revealing that it is in fact GPT-4o Mini. It did not think and wrote faster than 4o Mini.

`Environment`

Likely browser on Desktop (I got another user who was experiencing suspicious results to run it).

#
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.

flat knot
#

In addition to the three observations in yellow, the fact that it even complies on the prompt shows that it is not GPT-5. The prompt was crafted to be hard enough for GPT-4o and non-thinking models to fail every time, while GPT-5 Thinking can reliably refuse it.