I created a custom GPT, I want to access to it thru the API. I found that it’s not supported using the docs, I just do not believe. Why?
There is any workaround?
I see no one has answered. The answer appears to be no currently and for the foreseeable future. We have to pick and choose between assistants, custom gpts, custom models, and supplementing the base models real-time (fine-tuning, prompt engineering). It’s challenging to organize or even invest major refinement efforts in one thing knowing that we have to redo it for the other thing… Hopefully the product people realize that a lot of us are trying to demonstrate value in these products and the silos between the products are making that difficult.
GPTs are almost guaranteed to always be inside OpenAI’s walled garden.
If you want to use them, they must always be used through ChatGPT.
If you want to access them via API you must pay for it using the API endpoints available.