Integrate gpt40-mini

[guhlutions]

As the model "gpt4o-mini" is at least as powerful as gpt3.5-turbo but does not cost nearly as much, it should be usable by any API by now. Here the original quote from the OpenAI-Website concerning gpt4o-mini:

GPT-4o mini is our most cost-efficient small model that’s smarter and cheaper than GPT-3.5 Turbo, and has vision capabilities. The model has 128K context and an October 2023 knowledge cutoff.

[piscue]

I open PR in order to address this #53