docs: rewrite image reasoning example without multimodal LLM

[masci]

Description

Rewrite the "image reasoning" multimodal example leveraging the new ChatMessage and using the "plain" OpenAI class instead of the multimodal version.

Part of #15950

Before

openai_mm_llm = OpenAIMultiModal(model="gpt-4o", max_new_tokens=300)

image_documents = load_image_urls(image_urls)
msg = generate_openai_multi_modal_chat_message(
    prompt="Describe the images as an alternative text",
    role="user",
    image_documents=image_documents,
)

response = openai_mm_llm.chat(messages=[msg])

After

openai_llm = OpenAI(model="gpt-4o", max_new_tokens=300)

msg = ChatMessage(
    role=MessageRole.USER,
    blocks=[
        TextBlock(text="Describe the images as an alternative text"),
        ImageBlock(url=image_urls[0]),
        ImageBlock(url=image_urls[1]),
    ],
)
response = openai_llm.chat(messages=[msg])

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[logan-markewich]

Lgtm! I wonder if we want to note in the example that so far only openai multi modal supports this syntax so far?

[masci]

Lgtm! I wonder if we want to note in the example that so far only openai multi modal supports this syntax so far?

My only concern would be maintenance, i.e. remembering to update/remove the note as we roll out blocks usage to other providers...