Bug in models Gpt-o1 stuck in the "thinking process"

Hi, this last 30min i have a problem with the answer,
The model made the “thinking” but they cannot write the answer.
This happened at least 5 times in a lapses of 30-20min.
is wrote “Disseminated for 1m 20s” or “Finished Thinking” but nothing else.

The problem is general with the model o1 right now ? Model gtp-4 work without problem.
Thank you i advance.


If i try made the model answer me it stuck in the process.

Same problem here.

I tend to think that it happens when we exceed the limit of context in the conversation, but it should be handled in a clearer way by ChatGPT to understand what happens.

It happens to me with o1, o1-mini, o1 pro, changing to different models and activating or deactivating the “reason”, it can stay thinking for seconds, or minutes, and stays in different stages of thought (seeing the detail), but ends with “Finishing Thought”.

Often after 10 or 15 attempts it manages to give me an answer, but in general when I go from o1 pro to o1-mini or from o1 to o1-mini, so we really lose the advantages of this advanced model.

All this happens with long conversations, that is, with many previous messages, I have seen that it does not happen if I start a new conversation, but I would lose all the context so it would not be a viable solution.

In addition to handling this in a more clear way, I think ChatGTP could take advantage of previous “thoughts”, since it seems to restart on each attempt, quite similar to the previous attempt, but rethinking everything again, making unnecessary use of resources again, as well as surely consuming our o1 pro or traditional limits unnecessarily since we never get a response.

1 Like