How to Change Model Within a Project (folder)

Hi There,

Can you change the model of your chats(s) when you are with a Project (folder)?

The model selector is no longer in the top LEFT.

Thanks.

10 Likes

Welcome to the forum!

This is an excellent first question for those new to projects.

The current OpenAI ChatGPT FAQ for projects
Using Projects in ChatGPT

notes

Please note that chats in Projects use GPT-4o, and the model cannot be changed.

3 Likes

That’s a shame. I appreciate you taking the time to respond.

7 Likes

That is pretty strange and unintuitive

5 Likes

Yes this makes the whole Projects feature unusable which is a shame.

4 Likes

The workaround I’ve seen mentioned in the community is to start a new conversation with a different model and later move it into the projects folder.
I expect this restriction will be lifted at some point in the future.

Maybe it’s an option to manually select the model for the next reply as another workaround, like one can do in regular ChatGPT conversations?
My apologies if the answer appears obvious. I haven’t worked with the new projects feature, yet.

3 Likes

Yeah I just thought to myself “I’ll finally organise all these chats into a useful workspace”. Then I woke this morning to go and continue my coding work, only to find I could not change model.

I regularly work with with the basic model, but then, when it fails, switch to the o1 model to solve the problem, and then switch back. Since I can’t now switch back and forth within a chat, this makes the workspace feature unusable with my workflow.

Or is the idea that the user has different chats with different models in the one workspace, and each chat has access to the other chats in the workspace, so that the models are always in the context of that one workspace?

4 Likes

It’s using 4.5 now and I did not realize that until I got warning messages about running out of requests, when I got more in 7 days.

Really makes no sense to me - a “project” should just be a group of chats related to a topic, regardless of what model was used in each chat.

Just subscribing that it’s annoying that projects fix the model

ok, seems i was using a wrong client that didn’t expose it, it seems available now with the right client app. actually, the latest greatest mac desktop app still doesn’t support it seemingly

IT would seem you can now change the model at the project level. Which is great!

For me, when I change the model to, for example, o3-mini in a project chat, it moves out of the project and into the regular flow :face_exhaling:

1 Like

I’m having the same issue in the latest version of the mac desktop client. It works fine on the browser but it’s very annoying to not be able to select the model I want for the new chat from the project page.

Bumping this thread. I’ve also sent a support ticket but haven’t heard back. The issue seems to only affect iOS Projects, as desktop switching still works. But having to jump back and forth is super annoying and unintuitive. Adding a basic “change model” option inside the app would solve this instantly and shouldn’t be that hard.

Any ETA on a fix?

3 Likes

Having the option to organise chats into folders seems more useful than this project feature where you can’t change the model.

bumping the bump for support

Web version works for me now. Was previously having the same issues mentioned with the Mac OS app. I’m using Safari but I don’t think that would make a difference.

Are you sure about that? I can see the dropdown exists and will let me change it without popping it out of the project, but when I ask ChatGPT what model its using, it says its still using 4o.

Strange. I did the same thing and it was able to tell me when it changed models as well as the project name.

1 Like

Bump. This is still a problem for the MacOS app. I need to switch to the browser to be able to use a better model when I need to use a project’s context. Deepsearch an other tools are still unavailable inside projects on the app.

1 Like