Back to feed
News Story
Artificial Analysis (X)
1 sources

Google Releases Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS

Google has released Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS, its latest text-to-speech models supporting preset voices and voice replication. Gemini 3.8 Flash TTS debuts at #1 on the Pronunciation Robustness Benchmark and #2 on the Provider Voice Arena Leaderboard, while the Flash-Lite model ranks #6 in the arena and #1 in Japanese among multilingual arenas. The models are priced at $32.98 and $22.07 per 1M characters respectively, both faster than realtime.

SynthePulse Insight · AI deep readingMembers

Gemini 3.8 Flash TTS Tops the Pronunciation Chart, but Speed and Pricing Expose Trade-offs

Version 1 · 1 source

Google DeepMind has released Gemini 3.8 Flash TTS and Flash-Lite TTS. According to Artificial Analysis evaluations, the former ranks first on the pronunciation robustness leaderboard and debuts at second on Provider Voice Arena, but its speed still trails faster models such as Falcon 2, and both new models are priced higher than the previous generation Gemini 3.1 Flash TTS.

  • Google DeepMind has released Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS, supporting preset voices and voice replication.
  • Gemini 3.8 Flash TTS ranks first on the Artificial Analysis pronunciation robustness benchmark at 89.5%, with three Google models in the top four.
  • On Provider Voice Arena, Gemini 3.8 Flash TTS debuts at second with 1,263 Elo, and Flash-Lite TTS ranks sixth with 1,236 Elo, both above the previous generation's 1,199.
Open section navigationOne Release, Two Positions

One Release, Two Positions

Google DeepMind has released two text-to-speech models, Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS, both supporting preset voices and voice replication. This is the only fact in this release that can be confirmed as an official action; all other evaluation results come from Artificial Analysis's own benchmarks and leaderboards.

Artificial Analysis's evaluation places the two models in the same coordinate system for comparison: pronunciation accuracy, blind human preference, multilingual performance, speed, and price. Together these five dimensions sketch the outline of a trade-off—Google has secured a leading position in pronunciation accuracy, but has not simultaneously gained an advantage in speed and price.

Pronunciation Robustness: Google Takes Three of the Top Four Seats

On the Artificial Analysis pronunciation robustness benchmark, Gemini 3.8 Flash TTS ranks first at 89.5%, ahead of Gemini 3.1 Flash TTS at 88.2%, SpaceXAI TTS at 87.6%, and Gemini 3.8 Flash-Lite TTS at 87.4%. Three of the top four are Google models.

This dimension measures pronunciation accuracy and does not cover other speech quality dimensions such as naturalness or emotion. Therefore, the first-place 89.5% only shows that Google leads at "reading the words correctly"; it cannot be extrapolated to an overall speech quality lead.

Looking at the evidence chain, Google holds a leading position on the pronunciation accuracy dimension (this is an inference, based on occupying three of the top four seats). However, the source only provides a methodology link for the benchmark's specific test set and scoring method, without explaining them in the body text.

Free for now

Read the full analysis

5 more sections of analysis, plus the full takeaway

Loading

Credibility boundary

All evaluation figures in this article come from Artificial Analysis's evaluation post published on X and its own benchmarks and leaderboards, a single-evaluator source, not third-party independent verification. Google has not officially given the release date, available regions, or API access method in the source. The pronunciation robustness benchmark covers only pronunciation accuracy, not dimensions such as naturalness or emotion; the multilingual rankings cover only 9 Controlled Voice Arena languages; in the speed comparison, the source only mentions faster models such as Falcon 2 without giving their specific figures. Judgments marked as inferences in the text are all based on the rankings and price figures already listed in the source and do not represent Google's official positioning.

Sources for this version

  1. Google has released Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS. Gemini 3…Artificial Analysis (X)

Primary report

Artificial Analysis (X)

Primary source