Gemini 3.8 Flash TTS with Voice Cloning

Sam Witteveen covers Google's Gemini 3.8 Flash TTS in this 17-minute video. The first seven minutes go over the release itself, and the rest is hands-on in AI Studio and a Colab notebook linked in the description, alongside a Google blog post on the model.

After the intro, the video compares Flash TTS with Flash-Lite TTS at 0:53. Voice design comes up at 1:22, followed by a voice library of more than 2,000 voices at 2:44. At 2:58 he covers voice cloning and the guardrails around it. Stage directions and non-verbal tags follow at 4:02, then multi-speaker and long-form audio at 4:23.

Benchmarks take about two minutes from 4:48. Pricing is at 6:53; the description does not list the prices, so watch that chapter or check Google's pages for current figures.

The practical half begins at 7:19 with a demo in AI Studio. From 8:45 the Colab notebook takes over, with six chapters that each cover one feature, so you can jump to the one you need. It starts with a side-by-side of Flash and Flash-Lite. Style control is at 9:53 and inline tags at 11:05. At 11:37 the notebook returns to voice design, and multi-speaker dialogue is at 12:38. The last chapter, from 13:50, is voice cloning in Colab, so skip there if cloning is the feature you are evaluating.