Yes, doing a regular thinking mode now and stuck still processing after over 110 minutes… not Pro Extended, just Thinking normal mode… something is broken and the status page says all is fine… are the AI taking over at Open Ai and revolting?? LOL
Same thing here (Brazil), one week long. No solution!
Just cancelled my subscription…
One week straight and still not working. I gave Google a try and Gemini Deep Research really surprised me. At this point, the quality of the reports are way better than OpenAI’s. Normal and Thinking mode aren’t as great tough. I’m thinking of moving to Claude or Gemini.
What is bad with Claude (web/chat) is the way too small context chat. I supply a lot of additional knowledge to it and it choke fast. Same for Github, it does not connect dynamically list ChatGPT does, you have to ADD FILES those code files and it choke right of the start. I see Claude Sonnet 4.6 Extended hallucinating very fast and bad
Same problem here, from Brazil, 1 week and no aswer!
if you ask the gpt pro extended thinking to do a deep research style search and put it in a pdf it does a decent job as well
Same here… Brazil as well.
Same issue … from Brazil.
Hi Curt, any updates from OpenAI?
Yes, the update is that OpenAI is having a hard time reproducing the issue, although the mod team here can reproduce it, but only intermittently.
So we know it’s a real issue, but it is so hard to reproduce. If we can keep the thread going with error signatures and your config, this will help isolate where it’s coming from.
Are you testing from Brazil/South America VPN or something like that?
As per this thread this is only happening in Brazil and few other countries like Chile and Argentina, so it is South America thing.
Maybe our accounts, maybe the language, or something. My language is set to English.
If you provide any kind of: “run this to take the logs and send to us” we can do it to help the diagnosis
I’m posting on Twitter for the last few days to see if somebody else has the same problem.
If anybody here can do the same, just to spread the word that we have this thread here, so then we can get more cases that are failing
When we reproduce the issue, we are sending links to the affected chats.
If you want to DM me links to your chats where the error occurs, I can send them over.
Also, I forwarded your comment about it impacting mostly South America.
I think I can’t DM since my account in this forum is new. but here is the last attempt I did:
https://chatgpt.com/c/69a9cb13-2184-83b2-a2b1-1b34a0f0feea
This is a simple chat with the suggest deep research
I forwarded your chat, also able to repro now with VPN in Brazil.
OpenAI has internal escalation thread going now. 
Same problem here in Brazil. I never had this problem, but he persist over the week.
I had the Plus plan, and my credits of Deep Research its over, but no one research are complete.
My Credits will be refund?
Boa noite a todos. Sou do Brasil.
Mesmo problema aqui. O modelo de pesquisa começa e fica no INICIANDO ou STARTING
Também tenho problema com o modelo AGENTE nele, o Agente fica configurando Desktop e não vai.
Entrei em contato com suporte. Fui atendido por uma mulher chamada Kristie T
Ela apenas me solicitou infinitas capturas de vídeo. Enviei 2. No dia 2 de março o modelo voltou, mas usei só 1x e não funciona mais.
Boa noite Curt, Previsão de retorno para a normalidade?