Enumerates the supported LLM provider backends.

Used in ModelConfig to select which provider handles LLM calls. Each provider has a corresponding client factory and streaming implementation in the @hexos/runtime package.

enum LLMProvider {
    OpenAI = 'openai'
    Anthropic = 'anthropic'
    Ollama = 'ollama'
    Google = 'google'
    Azure = 'azure'
}