I'm really curious how long you've all experienced waiting for responses from the GPT-4o model. Personally, I've had cases where it took nearly **9 minutes** for a response, especially when my questions were very specific, like asking about medication contraindications without any additional links or images. I also noticed that tasks like image analysis can significantly slow down the response time; for example, it regularly takes longer when trying to find a video from a screenshot or identifying clothing styles from an image. As a Brazilian user, I wonder if the language used plays a role in this as well. So, what's the longest wait time you've encountered?
9 Answers
Around **5-8 minutes** for me when I asked about the missing cube puzzle with a screenshot attached.
I once waited **4 days** for deep research, but got nothing in return. It just kept saying it needed more time.
I waited **48 hours** once, and it didn’t generate anything. It just promised to provide something if I returned in 24 hours...but that didn’t happen.
The longest I waited was **27 minutes** when I asked some really detailed questions about a massive git codebase. I honestly didn’t expect it to take that long!
One time, it took **8 minutes** while I was frustrated with a Python project. I pasted a chunk of my code and explained what I needed. It spent 4 minutes just 'thinking' and then delivered 200 lines of code that mostly worked. After I clarified a bit more, it thought for another 8 minutes and gave me a second batch that ran perfectly!
All the numbers in your comment added up to 420. Congrats!
8 + 4 + 200 + 8 + 200 = 420
For me, whenever it hits **13 minutes thinking**, it just stops and shows an error like 'error in message stream.' The longest it actually ran without failing was **11 minutes** on a pretty tricky geometry problem.
Not too long for me, unless I’m doing deep research. That tends to extend the wait a bit.
I’ve had to wait about **2 minutes** once when I asked it about the role of merchants in Feudal Japan. I shared some of my thoughts and tried to draw some modern-day analogies, so maybe that added to it a bit!
Did you use the o3 model for that? I’m curious if it performed a lot of searches based on your question, since I think the response time depends on the complexity and depth of the answers it finds.
For me, it’s usually about **8-10 minutes** when I ask it to locate something from a screenshot. But coding questions typically get answered much quicker, like around 2-3 minutes.
Did it actually produce any results in the end?