Gemini 2.5 Flash and Pro TTS, Gemini 2.5 Flash Native Audio
2025-05-20 · Publisher : Google · Domain : Artificial Intelligence
At Google I/O, Google opens in preview the speech synthesis of Gemini 2.5 Flash and 2.5 Pro (two voices, more than 24 languages) and a native conversational voice for the Live API, improved in September and December.
What changes
- Expressive speech synthesis (TTS) with two speakers, capable of whispers, with smooth switching from one language to another.
- Live API with native audio output: more natural dialogue, control of tone, accent and style, tool calling.
- Experimental features: affective dialogue (emotion detection), proactive audio (ignoring background conversations) and thinking in the Live API.
- New native audio models in September 2025 (gemini-2.5-flash-native-audio-preview-09-2025) then in December (...-12-2025).
- December 10, 2025: improved TTS (Flash TTS for latency, Pro TTS for quality), precise expressiveness and pacing.
### Specifications - Identifiers: `gemini-2.5-flash-preview-tts`, `gemini-2.5-pro-preview-tts`, `gemini-2.5-flash-native-audio-preview-12-2025`. - Availability: Gemini API, Vertex AI; TTS in more than 24 languages. ### Pricing Current pricing from the official page (consulted in September 2026; the May preview pricing is not repeated): | Model | Input (USD / 1M tokens) | Output (USD / 1M tokens) | |---|---|---| | 2.5 Flash Preview TTS | 0.50 (text) | 10 (audio) | | 2.5 Pro Preview TTS | 1.00 (text) | 20 (audio) | | 2.5 Flash Native Audio | 0.50 (text), 3.00 (audio/video) | 2.00 (text), 12.00 (audio) | ### Benchmarks No numerical score published. ### Key takeaways Gemini's voice building blocks become usable for conversational agents; they remain in preview at the end of 2025.
Links
Tags : Audio, LLM, Google, Gemini, Voix