# LLMCapabilities Capabilities and features supported by an LLM service - `supportsChat` (`Boolean`): Supports conversational/chat completion format with message roles - `supportsCompletion` (`Boolean`): Supports raw text completion with prompt continuation - `supportsFunctionCalling` (`Boolean`): Supports function/tool calling with structured responses - `supportsSystemMessages` (`Boolean`): Supports system prompts to define model behavior and context - `supportsStreaming` (`Boolean`): Supports real-time token streaming during generation - `supportsSamplingParameters` (`Boolean`): Supports sampling parameters like temperature, top_p, and top_k for generation control [Java](../../java.md)