Hi why is o1 pro no longer generating any outputs? It’s completely broken right now. I’ve tried new chats, switching models. Nothing works.
6 Likes
I’m on the same boat. It hasn’t produced any answers for at least 6 hours.
4 Likes
same as everyone else, at least 8 hours doesn’t work
3 Likes
Same here. o1 Pro is not working for the last 6hrs. Please, fix the issue!!!
2 Likes
Same here. not working for me for what seems like weeks now. Its been on and off and now more off.
1 Like
same exact problem here, started about 9-10PM eastern time, so about 12 hours ago, wont output a single output, i was able to get a response after sending “hello” and that’s it.
please join me in sending email to support@openai.com so they become aware of the problem
1 Like
Same here, they busy with o3-Pro ?
Just saw an update to 1.2025.153 ?
Things look the same and the issue has not been resolved…
Tried another prompt, no output but just stuck now:
Been stuck like this for almost half an hour now. Stop button does nothing.
Did a new prompt, can confirm it shows an error now (See Gold’s post down below)
Same, tried new threads, nothing. I’ve had this happen on large threads in the past which was frustrating to have to start fresh with a new thread for a certain project. Now it’s happening with over o1 Pro thread for the past 14 hours.
Same. Mine was returning Korean characters in the details section last night. Hasn’t responded for the past week. And for the past two days all my models switch to 40 mini randomly and generate wrong or minimal response withing a few seconds.
same here in vancouver. not working for 6 hours already.
1 Like
I doubt it, considering its not a slew of users having this issues but select few, it must be openAI flagging select individuals and blocking their access
there is clearly something massively wrong
now it just says netowrk error occured, it takes like 7 minutes to reason on iPhone app, the longest of all, but still fails.
@Colm_Roche can you please check what is happening
1 Like
Submitted to support. I included links and screenshots. I wasn’t able to include those here.
Operator
Welcome to OpenAI’s Help Center!
User
Get started
Operator
How can we help?
User
Since about midnight last night, I haven’t been able to get any responses from ChatGPT o1 Pro. I’ve tried new threads, simple prompts, and reasoning, it producing nothing. I’ve cleared cache and cookies, etc. Still not working. I’ve seen others in the OpenAI developer community with the same issue.
Operator
Thanks for letting me know all the troubleshooting steps you’ve tried. There are no current outages reported that would affect ChatGPT or the o1 Pro model right now. To help further:
- Can you confirm if you get an error message, or does ChatGPT just remain blank after you send a prompt?
- Does this happen on all devices and browsers (if you’ve tried more than one)?
- Are other models (like GPT-4 or GPT-3.5) working for you, or is it just o1 Pro? Let me know and I’ll help you get this resolved!
User
Uploaded Screenshot 1:
The image shows a chat interface where a user asks about SQL assistance, and the system responds by first reasoning and then confirming SQL knowledge.
Operator
Thank you for sharing the screenshots—that’s very helpful. It looks like the model gets stuck on “Reasoned for 20 seconds” or “Confirming SQL Knowledge” in the “SQL Assistance” GPT, without ever giving an answer. A few more questions to help pinpoint the issue:
- Does this only happen with the “SQL Assistance” GPT, or does it affect regular ChatGPT (using Pro, GPT-4, or GPT-3.5) as well?
- Do you see the same behavior if you try a completely different GPT or a standard prompt in the main ChatGPT (not in a custom GPT)?
- Does it work if you log in from a different browser or device?
This will help narrow down whether it’s an issue with the “SQL Assistance” GPT specifically, custom GPTs in general, or all models. Let me know what you find, and I’ll guide you based on your results!
User
Uploaded Screenshot 2:
The image shows a conversation where a user repeatedly asks for help and guidance with a record labels table, but receives only messages indicating the system is “reasoning” or has “reasoned” without providing direct answers.
Operator
Thank you for the additional screenshot. I can see that the issue persists with other prompts and is not limited to just one GPT—it is happening with o1-pro as well, where the model “reasons” but never returns a response. This does suggest a deeper technical issue. To help further:
- Does this only occur with o1-pro, or does GPT-4, GPT-4o, or GPT-3.5 also get stuck for you?
- If you try using ChatGPT on the web from a different device or incognito window, does the problem still happen? Let me know the answers—this helps determine if this is model-specific or account-specific, which is important for resolving your issue.
User
Happens on all devices and browsers. Mainly appears to be impacting o1 Pro.
Operator
Thanks for clarifying that the issue happens across all devices and browsers, and mainly impacts the o1 Pro model. Other users have also noticed similar issues, so this may be a problem specific to the o1 Pro model right now. Here are a few quick steps you can try while this is being looked into:
- Switch to GPT-4, GPT-4o, or GPT-3.5 models for now (these should still work normally).
- Try using o1 Pro again later—sometimes model availability or performance can be temporarily impacted. If the issue continues with o1 Pro after some time, please let me know here with any updates. Thanks for your patience—your reports are very helpful for monitoring and fixing these kinds of problems!
User
Here are 2 threads of others having the issue that are all from today, and all o1 Pro Mode:
o1 pro is partially down multiple reports here all pro accounts
o1 pro stopped creating any output
Operator
Thank you for providing those community threads. I see this is affecting multiple users with Pro accounts, specifically when using the o1 Pro model. Your details and links are very valuable in tracking and understanding the scope of this issue. For now:
- Please use GPT-4, GPT-4o, or GPT-3.5 as a workaround—they should still respond normally.
- If you notice changes (for better or worse) in o1 Pro’s behavior, feel free to update me here. This does appear to be a known, model-specific issue and is likely being investigated based on the number of reports. Thanks for your patience while it gets resolved! If you have any other questions about using ChatGPT or need any support with other features, I’m here to help.
My friend subscribed to pro yesterday and all the models were failing, so they gave her a refund. Even if it is selected accounts, it seems to be random.
Definitely an issue. Been happening for >24hours
Edit: might also be worth noting that I only noticed this because o3 has been significantly worse in the past few days (truncating abruptly, not finishing output) to the point of being unusable.
I’ve had to resort to using gemini for long outputs
where is our refund, I want $6 back for this outage
did you send an email to the support?
I have never been able to solve a single thing with o3. Anyway, still no o1pro working. I noticed they released some new audio feature? So what they just burnt some of their pro user compute to facilitate it? Openai seems shady at this point.
1 Like
The same here it does not work
She was a new subscriber, her account was failing constantly so they issued a full refund. You can request a full refund if you explain the issues but that means your account will be deactivated. On the other hand, support is not responsive or a machine might respond back to you. When they escalate the issue and how, I do not know since I have been dealing with this since third week of May. Today I got an email from OpenAI that stating “o1 pro mode is now limited to 5 queries per month per user due to its computational intensity.” I am assuming the bot might have made a mistake as this is not written anywhere for pro users, or this is hidden and not disclosed to us yet! So even emailing them so far hasn’t helped but has caused more confusion for me.
advanced voice, dallee, sorana or whatever the hell video generator, are all a massive waste of resources, i disabled advanced voice immediately when realized it was giving short answers, but it has also broken regular voice chat to the point you cannot use it anymore, i literally have to use voice recording, where it will transcribe, send as text, then playback answer, im a verbal strategist so need voice but not stupid advanced voice, gimmick for kids