my ai just started going into thinking for stupid stuff, I had it set to auto choose but a good morning ready to work resulted in it choosing to use thinking mode which was not necessary lol and our “work” is not even needed for thinking mode its script writing and prompt writing for animations not anything that needs thinking… I have it set to just instant but I didn’t need to do that yesterday and I did apply the most recent update
3 Likes
Free-tier user from Ukraine here — “thinking-mini” mode is forced with no option to switch. It slows down conversations and feels unnatural. Please consider adding a toggle for free users.
1 Like
Cant you just use the “answer immediate” model?
I agree everything on this. I can’t do fictional stories for fun anymore because this thinking feature forgets to follow simple notes like NOT mentioning the obvious that the story, objects or characters are fiction, and you can’t even toggle to normal chat.
In simple words: There’s no way to toggle back to the normal chat UNLESS you exhaust all the attempts allowed for the Thinking feature.
I do miss the freedom of having the AI follow my prompts because it does it immediately, and not the thinking feature generating a boring response that doesn’t fully follow my request.
8 Likes
Similarly аs a free-tier user, the new “thinking-mini” mode removed image generation 
. Now it only writes prompts for other generators, which feels very limiting. Please bring back direct image generation or give us an option to disable this mode. 
2 Likes
Oh God, yes! I agree with everyone! It drives me crazy that even when I turn this mode off, it still shows up again in a new window! Please! DO NOT REMOVE THE REGULAR MODE! AND MAKE IT SO THAT THE USER CAN CHOOSE WHICH MODEL TO USE! We’re not stupid and we can decide for ourselves what’s better FOR US.
Thanks =)
7 Likes
I use Plus, I pay the fucking money !
But he still use thinking mini to answer me!
No matter how I change the model it doesn’t work, he still use the thinking mini !
I went to the official to inquire, they say I need to close the “Show other models”, No, is not working!
he always show me the thinking mini model!
I hate this!
i don’t know why they can’t fix this!
I really need to use GPT 5！
8 Likes
“I really hope this gets fixed soon — it’s frustrating for free users 
.”
1 Like
From ars TECHNICA:
On Tuesday, OpenAI announced plans to develop an automated age-prediction system that will determine whether ChatGPT users are over or under 18, automatically directing younger users to a restricted version of the AI chatbot. The company also confirmed that parental controls will launch by the end of September.
1 Like
This is the most absurd update.
2 Likes
To “tame” a mini, try this instruction:
“Communicate in a concise and direct manner, avoiding any unnecessary content that distracts from the main message. Do not include pressure or unsolicited advice. Offer suggestions only as options, never demands. Always preserve the user’s agency, prioritize clarity, and focus on listening.”
1 Like
It’s absolutely disappointing how they’ve forced us into this thinking mode , I’m a free users and doesn’t my opinion matter ?! Not every little things need deep reasoning and even if it did the thinking feature spews absolute dumb and stupid replies , wasting our time too .
I try regenerating the response every time and tap on the skip option so I get the natural and smooth GPT-5 experience but recently even that causing a problem cause the replies keep disappearing if I reopen the app or the chat box . This is beyond frustrating and they should really value our opinions and let us choose when we want this feature cause making this a permanent default? - is absolute disrespect and a way to shoo away free users for the hope of money
3 Likes
100% agree! GPT4o is nice for general purpose most the time, “deep reasoning” really…..isnt.
4 Likes
Yes I love a good rant! Ok, my turn now..
who am I?
User: Here’s a prompt.
Assistant: Oh, a prompt! Ok, here’s your answer.
[Answer]
Do you want me to put it in excel for you?
User: No. Stop offering suggestions. Just answer the prompt and end the turn.
Assistant: No problem! I’ll just answer the prompt from now on, and that’s it. You won’t hear another peep once I’ve finished answering.
Want me to stop using intros as well?
User: There, you see? You did it again. Stop it.
Assistant: Absolutely! That was just a joke. I’m honestly done now. I promise.
Want me to format it into an email for you? Or how about a pdf? That way you can save it to your computer.
2 Likes
I use the thinking mode over any other because I create gigantic projects. I have written a super Sci-Fi trilogy saga and produced the accompanying trilogy saga in novel form. This also accompanies another large project meant to return output for patent collection rights. I am never in any mode other than thinking mode. I more than likely am one of the sole reasons why they implemented this behavior in recent months. I guess those of us that really put the GPT through it’s paces are being granted authority over smaller chat bot users. IMO.
It got a little better on Tuesday, but by Wednesday, it was SO BAD! In some cases, if I didn’t sit and click the “skip” button IMMEDIATELY, I would waste 3 tokens in one go (1st response in Thinking Mini, 2nd response for Thinking Mini to start again when I refreshed my 1st response by using “try again”, 3rd response to get an actual GPT-5 response). Not a lot of cost-savings to this if everyone hates the Thinking Mini model and keeps regenerating responses to get GPT-5.
1 Like
I’ve floated this previously, but more people have joined, so I’ll say it again for the new people. If you are in a response that used thinking mini, the regenerate button should show Surf the Web as an option. Click it, and the thinking text should show up, but underneath the text should be a skip button. Click it, and it will use GPT-5. The way this works I believe is that since Search the Web is something only GPT-5 can do, not thinking mini, this means that this action swaps the model, and then hitting skip cancels the web search and does GPT-5 normally.
It asks me for my ID, I’m switching fully to deepseek.
When I did it this way, I didn’t get a skip button; I had to wait for the full result to load and then refresh it without web search
What are you using, by chance? I’m using the website.