Gemini 2.5 Pro TTS

What it does

Gemini 2.5 Pro TTS takes written text and returns spoken audio. Its only capability on Pipe2 is text-to-speech, so it accepts a script and produces a voice track. It does not take audio in, transcribe, or generate music. It is described as studio-grade on prosody and pacing, which points it at production narration rather than short utility blips. You reach it through the Audio Generator pipeline.

When to use it

• Running Audio Generator on a narration script (voiceover for a video, a course module, an explainer) where sentence rhythm and pauses carry as much as the words.
• You need a specific delivery: Audio Generator accepts custom style directions for tone, accent, and pacing, and this model is described as studio-grade on both.
• You are producing in a non-English language; the Audio Generator surface covers 70+ languages.
• You want to audition several readings of the same script, since the pipeline exposes dozens of voices behind one text input.
• Skip it when you only need a short system prompt or notification line. The pacing control is wasted on one clause.

What to know before you run it

The wholesale unit for this model is quoted in output tokens (roughly 1500 per unit), so the amount of speech the model produces is what the unit measures. The live price table on this page shows the current rate for the pipeline. You steer the read with the text style directions Audio Generator exposes for tone, accent, and pacing, so plan on a couple of passes to land the delivery you want. This model has no published examples on Pipe2 yet, so preview a short excerpt before committing a full script.

Also known as

You may see it written as Gemini 2.5 Pro Text-to-Speech, Gemini 2.5 Pro Preview TTS, Gemini-TTS, or under the preview identifier gemini-2.5-pro-preview-tts. These all refer to the same model listed here as Gemini 2.5 Pro TTS.