A field engineer wants to submit a photograph of equipment and ask a question about visible damage. Which model capability is necessary for a direct image-and-text request?

1. A text-only embedding model with no image preprocessing or image support.
2. A model that supports the required image input and text question modalities.
3. Any model with the largest output-token limit, regardless of its supported inputs.
4. A speech synthesis model that outputs audio from text.