Low-Latency Private Knowledge Chat: Choosing Streaming Models for Healthtech

Short answer: Use GPT-4o-mini for the latency-first path: its streaming time to first token is about...