Observation: Lower Response Quality in "Projects" Mode Compared to Normal Chat (GPT-4o)

ChatGPT
, ,
1

Hi everyone,

I’ve been testing the “Projects” feature and noticed a consistent difference in response behavior that might interest others.

Observation: When I start a project without any initial instructions or files, the model (GPT-4o) tends to produce shorter, less detailed, and seemingly lower-quality answers to the same basic prompts compared to a regular chat session using the same model.

In addition:

  • Vision capabilities (e.g., analyzing images) also feel less reliable in Projects. I encountered more hallucinations and misinterpretations in image analysis when using GPT-4o inside Projects, versus much more accurate results in regular chats.

This makes me wonder:

  • Is the context behavior of GPT-4o somehow limited or sandboxed in Projects?
  • Are there model configuration differences we’re not aware of?
  • Or does the lack of initial instructions trigger a default behavior that’s more conservative or minimal?

Would love to hear if others have experienced the same.
Any insights from the OpenAI team on this behavior?

Thanks!

1 Like
2

I have just started using projects to try to organise my chats, but I googled this and found your thread because I have found the same thing and I wondered if anyone else had experienced it.

To clarify, I find the responses significantly worse in project chats than in other chats. The project chats seem to answer my questions with this constant structure that is not helpful for what I’m doing. The model often doesn’t answer my full question or consider everything I’ve asked, and I find the information to be explained less well.

I think I’ll be going back to normal chats.

What have you noticed specifically?