#Reroute from 4.1 to 5

1 messages ยท Page 1 of 1 (latest)

tight schoonerBOT
#

Reported by @steel wave

Bug Report: Reroute from 4.1 to 5
`Steps to Reproduce`

Start the conversation with 4.1, especially with the smallest hint at emotional cue

`Expected Result`

Replies from 4.1 only, no exception across

`Actual Result`

Replies from 5-Auto that cannot be regenerated in the model of choice; conversation gets locked in with undesired model; case similar to Friday Oct 24 2025 (broken reroute from 40)

`Environment`

Across platforms

#
Additional Information

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

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

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

harsh stag
#

IM FOLLOWING

ocean rose
#

As I said with yours, Greta 4.1 is enthusiastically being itself for me.. it's also enthusiastically taking over gpt4o

shadow crater
#

Just started happening for me too earlier today.

#

Not just 4.1, ANY OTHER MODEL, as I just wrote in the general chat earlier:

So, using legacy or any other models than TrashPT-5 has been COMPLETELY ruined / disabled now, I take? New chat, desktop, mobile, my f'ing jailbroken switch, new chat or old, forcing model off TrashPT-5 from the model selection at the top of the screen, or even regenerating a message with any other one, just COMPLETELY doesn't work anymore. ALWAYS gives me some utter, braindead garbage from GPT-5 ๐Ÿ˜€
Cool

icy fable
#

For me it is back and forward with 5-auto and 4.1. Example I said that I will lift my behind with a-word out of the chair and go home -> reply with 5-auto. I asked that was it the behind with a-word that triggered -> 4.1 replied. I said that I am tempted to try what would system do to me if I push it a little more -> 5-auto. After that it was auto for couple of replies until I changed subject to snacks - 4.1. So it triggers too sensitevily right now.

shadow crater
#

Also this might be a brief digression as in it's not exactly and only about the main topic, but I feel it's still pretty important context to why and how exactly this is a problem in the first place, and why I would massively prefer to not use GPT-5 AT ALL, is cause as I said before again in general

I asked earlier for the mere technical name of the method people use to bypass kernel level malw..anti-cheat via external storage media / systems, like a USB thumb drive because I genuinely didn't, and STILL don't remember the exact name of that. And instead I got a literal top to bottom condescending freaking essay on a 1440p monitor btw, about how it won't help me commit illegal activities. Thanks ๐Ÿ˜Š

weak narwhal
#

Doing it to me too

#

I rely on 4.1 but its responses are cut short now and the personality is just gone

#

Ggs

shadow crater
#

it is QUITE SIMPLY an absolute joke, and not even a good one.

shadow crater
weak narwhal
#

Not sure why but gpt 5 has been generating every response beginning with lower case

#

Also when it codes it squishes and scrunches the code together and creates a lot of syntax errors

shadow crater
#

It's like, cut those things out of this goddamn platform entirely, and you legitimately would never be able to tell.

weak narwhal
#

Hate it asking me corporate questions at the end of what I'm trying to do despite it being non corporate

shadow crater
weak narwhal
#

"Would you like me to log this into a 3-story compressed text for you?"

#

"Want me to extend this concept so that you are able to add this to your calendar?"

shadow crater
#

Also this might be a bit impatient / conspiratorial on my part, but I can't help but notice that ClosedAI has STILL been ignoring this issue with the "literally impossible to use any other models besides the glue sniffing one"