I developed a Custom GPT using GPT4 that is able to receive images as inputs and interpret them. The GPT is working exactly as planned.
I am trying to replicate the custom GPT with assistants so that I can use it in a third-party app. The problem is that I am not able to find an Assistants GPT model that is able to receive and view images as inputs.
I realize that the gpt-4-vision-preview API is able to solve my issue but I am not an experienced developer and would much rather use Assistants tools to build GPTs than have to develop the API in Python.
When will the gpt-4-vision-preview API become available to use for Assistants?
A GPT is the worst name one could choose ever, and is meant to represent the enhanced shareable plugins available exclusively within and by ChatGPT Plus.
Assistants are also not a great name, but are an API agent that remains under the control of the API developer.
Being an agent, the power comes from being able to make recursive calls autonomously to utilities such as a knowledge browser, a python code sandbox, or to the API developer’s own functions.
I can see vision being problem-fraught in an agent scenario. The best case I could come up with is to not need to have your API switch to an image in → description-out-only scenario when a user provides a picture. “look at these pdfs” could be uncontrolled billings.