Your team routes all LLM calls through a new central gateway that runs a routing classifier on each request. Median latency for a short autocomplete feature rises noticeably. What is the most likely reason?

1. A Gateways add no measurable latency, so the provider must have slowed down that week.
2. B The gateway forces every response to stream, which is slower for short completions.
3. C The gateway adds a network hop and classifier time before the provider call.
4. D The gateway's prompt cache misses on every request, which doubles model latency.