Google Teases New Gemini Flash Text-to-Speech Models

An announcement attributed to a Google figure claims new Gemini Flash and Flash-Lite TTS models with thousands of voices, cloning, and broad language support, though the news comes solely from a social post so far.

A post from a Google-affiliated account on X introduced what it called Gemini 3.8 Flash and Flash-Lite TTS, pitched as new state-of-the-art text-to-speech models. The announcement lists a revamped voice design workflow, a library of more than 2,000 production-ready voices, and the ability to clone or replicate voices directly within the system. According to the post, the models cover 100 languages out of the gate, positioning them as a broad-reach option for dubbing, narration, and other voice-driven production work. The same message also teases a forthcoming "voice remixing" feature, suggesting Google plans to let users blend or adjust voice characteristics in a later update rather than at launch. The post further claims the new TTS system has taken the top spot on Hume AI's voice benchmarks, though no benchmark scores, methodology, or third-party confirmation were included. As of now, this single social media post is the only available account of the release, with no accompanying blog post, documentation, or pricing details to verify the specifics. Given the scale of the claims, particularly the voice count and multilingual support, the AI video and audio production community will likely be watching for an official Google rollout or developer documentation to confirm access, availability, and real-world performance before the tool sees wider adoption.