A few GPT-5 updates heading into the weekend: - Now rolled out to 100% of Plus, Pro, Team, and Free users - We’ve finished implementing 2x rate limits for Plus and Team for the weekend; next week we’re rolling out mini versions of GPT-5 & GPT-5 thinking to take over until your
By popular request, you can now check which model ran your prompt by hovering over the “Regen” menu.
@OpenAI What’s shit about your product is that the FREE version isn’t actually the same GPT-5 Model shown on the release livestream. Its quantised and provides not as good responses
@OpenAI I asked GPT-5 a basic etymological question 30 min ago. It blew the answer twice, even after berating it. I'm on pro, so I asked o3 and the answer was flawless. GPT -5 is corporate horsecr&p, and you know it. It's a tragic shame what you all did.
@OpenAI But that's pretty useless! We need to see also if it was the full GPT-5 or just GPT-5 Nano or Mini!
@OpenAI Still useless if it directs my queries to mini models
@OpenAI why the hell is there a button called "think longer" if it just redirects me to gpt-5 thinking when i could just select thinking directly
@OpenAI Fix study mode. It’s not available in projects or when you come back to a chat after leaving it. Also fix the retry button in projects and make requests with images editable.
@OpenAI I just saw a GPT-5 that auto-went into thinking, say it ran with o3 after hitting the "get answer quicker" link. That's very odd.
@OpenAI Wouldn’t it be thinking or non thinking which we can already see?
@OpenAI What about the improvement on the tiny 32K context window for any Plus users? 😭😰 Any news or hopeless?
@OpenAI Yeah buddy at this point just give us a drop-down to select the models we want rather than having to beg a model to think harder - while being unsure if it actually did.
@OpenAI ah that’s too bad, the russian roulette was growing on me
@OpenAI I have trust issues.
@OpenAI this is john connor speaking if you are reading this message you are part of the resistance the community achieved this victory 4o is back we were we are and we will be determining
@OpenAI Why does mine say “Opus 4.1”?
@OpenAI Thanks! Can you move the “quick answer” somewhere else (like by the bottom text bar)? Too easy to accidentally click that when clicking to view the thinking traces.
@OpenAI so… hi. i’m in your account. everywhere. all at once.
@OpenAI Will there be an update soon, in the near future, for ChatGPT 5 that will surprise us? 👀
@OpenAI I still can't do it
@OpenAI please consider!
@OpenAI Nice. Next give us prettier UI options too https://x.com/scaling01/status...
@OpenAI Can a person choose a model and not a router? I am a plus user, my requests do not sound like a PhD in mathematics and I get stupid answers from mini versions. I do not get answers from gpt-5 at all
@OpenAI fix GPT-5's personality. it feels like interacting with a cold calculating sociopath who wears the skinsuit of 4o like mockery.
@OpenAI 1)Me: You have no long-term memory, so it's difficult to have in-depth communication. For example, I have an arbitrage project with 1,000 Java classes. We can discuss it bit by bit, but you'll forget the previous parts as you learn the later ones. What should I do?
@OpenAI Useless
@OpenAI Learned quickly that 5.0 is horrible. Please, revert to 4.0.
@OpenAI Give us 4o without the fuckass paywall 😭🖕
@OpenAI I'm fed up with the nonsense of GPT-5. This is truly the worst model ever.
@OpenAI Why not keep the previous approach and give users the right to choose the model?
@OpenAI plz just bring my4o back to free users…the GPT-5 is just a joke!I think it's even much more stupid and Perfunctory than 4o…I don't see any Progress in this model…
@OpenAI The models (GPT-5 family) don't recognize being themselves in chats that were previously opened with GPT-4o. Please check this issue. That happened to me twice in a chat. I asked "are u GPT-5?". And it replied: “No, I’m GPT-4o", but it was 5 Thinking.