Stop event from a streaming response.

Emitted when the stream ends. Contains the reason generation stopped and optionally the final usage statistics.

Fields

• :reason — why generation stopped:
  • :stop — natural end of response
  • :length — hit max_tokens limit
  • :tool_calls — model wants to call tools
  • :content_filter — content was filtered
  • :max_tokens — provider-specific max tokens reached
• :usage — final LLM.Usage.t() with token counts (may be nil)

Example

%LLM.Stream.Stop{
  reason: :stop,
  usage: %LLM.Usage{input_tokens: 10, output_tokens: 50, total_tokens: 60}
}