GPT Plus: ChatGPT has become more stupid and always uses the 4o mini model for responses

ChatGPT Bugs
,
1

– Main problem: ChatGPT functions do not work properly anymore and use the worst - 4o mini - model for its responses –

Please get a quick fix, or tell when it may due - 4o mini is really unbearable in terms of quality of answers when compared to 4o, 4.5, o3, etc.

To check which model is used for the response:

  • 4o tends to reply with emojis (without prior settings)
  • (just an observation from me, could be just me) the responses from 4o mini are much faster printed compared to the more advanced models.
  • 4o and all subsequent models are able to count letters of a long word (ex. “how many 'r’s are in strawberry?”)
3 Likes
2

I have the same issue. I subscribed to GPT-Plus, and since yesterday, even though I choose to use the 4o model, the system responds using the mini model without any prompts. (My 4o hasn’t exceeded the 3-hour limit.)

3

It’s unfortunate that even though we paid for this service, the OpenAI team can’t even get something this mundane right.

Do you think it could be our browser settings? Are you using a VPN? Or have you shared your account?