Hi everyone,
I’m working on a project where we handle a lot of data, and we’re trying to use this data in a custom GPT. However, we’ve noticed that the GPT sometimes gets the data wrong or makes up information. We’re considering using the new Assistant API v2 because it seems better for managing lots of data.
Does anyone know if it’s possible to combine an Assistant API with a Custom GPT? This way, we could have the robust data handling of the Assistant API with the user-friendly interface of the Custom GPT. Any advice on how to do this, or any workarounds you know of, would be really helpful.
Thanks for your help!
2 Likes
Looking forward to seeing the answer here as I’ve been wondering the same.
If not, maybe GPT5 updates will bring a larger knowledge database for custom GPTs…
Yeah, I hope there is a solution for it.[quote=“highland.hllc, post:2, topic:735579, full:true”]
Looking forward to seeing the answer here as I’ve been wondering the same.
If not, maybe GPT5 updates will bring a larger knowledge database for custom GPTs…
[/quote]
That is a very quick win I guess, using the new Assistants v2, GPTs can have better retrieval. I hope that with gpt5, the interpreter environment is also more flexible & is enabled to create teams of assistants working on a project. better accuracy is a prerequisite for planned steps though since (1 - error_rate)^steps = p(end state)