What is Time to first token?

How long until a model starts producing output.

Time to first token measures the delay before a model emits its first piece of output rather than the time to finish. On a streaming voice call it is the figure that governs perceived responsiveness, because speech synthesis can begin before the full reply exists.

This is an industry glossary. A term defined here is not a statement that Drop AI provides it — see the pricing page for what is included.