I want to measure end to end inference speed with a longer English prompt so the engine has to process more context and maintain attention over time. Please synthesize clear, natural speech with steady pacing and expressive intonation, as if you are calmly explaining a technical topic to a friend. Avoid dramatic pauses, but keep the rhythm varied and human. When you finish the last sentence, stop cleanly without trailing noise or extra words. This prompt is intentionally about one hundred words long to provide a stable benchmark across runs.
