Time to first token
Also called TTFT
How long you wait between sending a request to a model and the first piece of its reply arriving.
A request's time splits in two. First the model reads your prompt, all of it in one pass, which is usually quick. Then it writes the reply one token at a time, which is usually most of the wait.
Time to first token is the first part plus the first step of the second, and any thinking the model does before it answers. It decides whether a feature feels responsive; streaming is what lets the reader see it.
For example
A support assistant that streams its answer shows the first words in about a second, while the rest of the reply keeps arriving for several more.
Put it to use
Where you'll meet time to first token in ChatGPT, Gemini and Claude, how to use it there, and the mix-up most people make.
Daily Bytes Plus is in the Android app.