I have been testing out LLMs with the together.ai API, but I can't figure out how to use the multimodal models with the API. I don't see any in their model list.