I would like to suggest the implementation of a feature that enables the model to perform real-time searches when it cannot find information within its existing knowledge base. This is particularly important when using ChatGPT-4 or ChatGPT-4o, as the standard GPT-4 allows for such functionality to avoid situations where answers are unavailable due to limitations in the uploaded knowledge base.
Thanks!
