Voice · Google DeepMind

Gemini 3.1 Flash TTS

Google's April 2026 voice model, steered with tags like [whispers] written into the text. Pick it when delivery matters more than price.

Anonymized Very high cost
Latest versionGemini 3.1 Flash TTSApril 15, 2026
Artificial Analysis Speech Arena1,209 Elo#3 of 10 on Artificial Analysis Speech Arena
Price$0.0843 per 1,000 characters$0.253 a page read aloud
Key spec70+ languages

What sets it apart

Gemini 3.1 Flash TTS lets you direct delivery in plain text with audio tags and Director's Notes, and covers more than 70 languages. Google has since moved the family to Gemini 3.8, so this is the previous generation.

  • Audio tags and Director's Notes
  • 70+ languages, multi-speaker dialogue
  • SynthID watermark on all audio

Recommended for: a voice that acts: whispers, laughs, pauses.

Strengths and limits

Where it shines

  • More than 200 audio tags for tone, pace and accent
  • 70+ languages
  • 30 voices

Think twice when

  • The dearest voice in the catalog per character
  • Anonymized, not private
  • 30 voices
  • Up to 5,000 characters per request in Sub Rosa

Use cases

Two-person podcast dialogue

Speaker 1 [excited]: We just launched! Speaker 2 [whispers]: Don't tell anyone yet.

Tags give fine control of mood per line.

Multilingual course narration

Narrate this lesson slowly and clearly, with a friendly teacher's tone.

Wide language coverage and steerable tone.

Against its rivals

  • ElevenLabsElevenLabs is cheaper per character with fewer languages and less tag control.#1 of 10 on Artificial Analysis Speech Arena (1,330 Elo)
  • xAI TTSxAI also uses inline tags, at about a tenth of the price.
  • Inworld TTS-1.5 MaxInworld is far cheaper and expressive, with 15 languages.
Compare side by side ↗

Version history

Each version, dated, with what it changed over the one before.

Artificial Analysis Speech Arena, version after versionElo rating · Artificial Analysis

Hollow points are earlier versions no longer offered in Sub Rosa.

See the numbers
VersionReleasedElo
Gemini 3.1 Flash TTS2026-04-151,209
Gemini 3.8 Flash TTS2026-09-221,275
  1. Gemini 3.1 Flash TTSLatest
    • Introduced granular audio tags for style, pace and delivery
    • 70+ languages and native multi-speaker dialogue
    Source : blog.google
  2. Gemini 3.8 Flash TTSEarlier, not in Sub Rosa
    • Generally available, over 130 languages, inline tags like <sigh> and <breath>
    Source : ai.google.dev

Benchmarks

How to read these scores: what a benchmark tells you.

Artificial Analysis Speech Arena: #3 of 10 models measured in the catalog. People listen to the same sentence read by two voices, without knowing the models, and pick the more natural one. An Elo rating: higher sounds more natural to more people. 100 points apart means the higher-rated model wins about 64% of duels.

Artificial Analysis Speech ArenaElo rating · Artificial Analysis, October 4, 2026
  1. 1. ElevenLabs TTS v4 TurboElevenLabs1,330
  2. 2. ElevenLabs TTS v4ElevenLabs1,329
  3. 3. Gemini 3.1 Flash TTSGemini 3.1 Flash TTS1,209
  4. 4. ElevenLabs TTS v3ElevenLabs1,178
  5. 5. Gradium TTSGradium TTS1,151
  6. 6. MiniMax Speech-02 HDMiniMax Speech-02 HD1,114
  7. 7. Chatterbox HD (Resemble AI)Chatterbox HD1,102
  8. 8. ElevenLabs Turbo v2.5ElevenLabs1,097
  9. 9. ElevenLabs Multilingual v2ElevenLabs1,095
  10. 10. Kokoro Text to SpeechKokoro1,064
See the numbers
ModelElo
1. ElevenLabs TTS v4 Turbo1,330
2. ElevenLabs TTS v41,329
3. Gemini 3.1 Flash TTS1,209
4. ElevenLabs TTS v31,178
5. Gradium TTS1,151
6. MiniMax Speech-02 HD1,114
7. Chatterbox HD (Resemble AI)1,102
8. ElevenLabs Turbo v2.51,097
9. ElevenLabs Multilingual v21,095
10. Kokoro Text to Speech1,064

The model in this family

ModelUsePrivacyPrice
Gemini 3.1 Flash TTStts-gemini-3-1-flashVoiceAnonymized$0.0843 per 1,000 characters$0.253 a page read aloud

Specifications

SpecificationGemini 3.1 Flash TTS
ReleasedApril 15, 2026
WeightsClosed
Acceptstext
Producesaudio
Languages70+ languages
Price$0.0843 per 1,000 characters
PrivacyAnonymized

Prices, context and lengths from the live catalog on October 16, 2026; parameters and licenses from the maker.