# What is time to first token (TTFT)?

Time to first token (TTFT) is the delay between sending a request and receiving the first token of the response. It is the latency number users actually feel in streaming interfaces, and it is dominated by prompt processing, so it grows with input length while decode speed governs the rest.

TTFT and total latency answer different questions. A chat interface with fast TTFT feels responsive even while a long answer streams; a batch pipeline only cares when the last token lands.

The practical levers on TTFT are shorter prompts, prompt caching (a cached prefix skips most prompt processing), and routing interactive traffic to models sized for it.

## Related terms

- [Inference latency](https://allocate.network/glossary/latency.md)
- [Tokens per second](https://allocate.network/glossary/tokens-per-second.md)
- [Prompt caching](https://allocate.network/glossary/prompt-caching.md)

---

[HTML page](https://allocate.network/glossary/time-to-first-token) · [Machine-readable catalog](https://allocate.network/catalog.json)