AI SDK embedding requests use a language model and fail before sending #13370

Embedding requests through the AI SDK adapter fail with TypeError: model.doEmbed is not a function.

On main (5522c6f), construct AiSdkApi with model: "openai/text-embedding-3-small" and call embed({ model: "text-embedding-3-small", input: "hello" }). The adapter gets a language model from the provider's callable interface and passes it to the SDK embedding function. The request fails before reaching the endpoint.

I reproduced this locally with the installed AI SDK. As a control, provider.embeddingModel("text-embedding-3-small") sends the expected /embeddings request and returns the embedding using a mock fetch response.

I plan to select the provider's embedding model with the configured model ID and add single-input and batch regression tests.