Use OpenAI vision

[AvivSpiritt]

Hi,

I'm trying to use OpenAI's vision capabilities, but I'm unsure how to pass an image using image_url.

When I tried using it, it seems the model treats image_url like any other text.
(not sure if I can pass an object instead of text)

  (generate
   [[:system "You help identify images"]
    [:user [{:type "text"
             :text "Whats in this image?"}
            {:type "image_url"
             :image_url {:url base64-image}}]]
    [:assistant (llm :openai-azure)]])

Thanks 🙏