Text to embed.
The model that produces the vector. Which names are valid depends on the server's provider.
Model-specific options, passed to the inference service as they are.