Google Gemini 3.8 TTS: Can replicate voices from just a 30-second sample

1 days ago

Beating AI Express (Dongcha) reports: Google has launched Gemini 3.8 Flash TTS and Flash-Lite TTS, two text-to-speech models now available on Gemini API and Google AI Studio. Flash TTS focuses on audio quality, character performance, and long-text stability, while Flash-Lite TTS is tailored for high throughput, low latency, and low cost. Both models allow users to design new voice lines directly from text, or replicate personal or authorized voices using a roughly 30-second reference audio clip. Users can control emotion, rhythm, and accent on a sentence-by-sentence basis, and insert sounds such as laughter, sighs, or coughs. They can also generate two-person dialogues directly from a single script. Flash TTS supports 130 languages, while Flash-Lite TTS supports 101. Pricing is lower than the previous generation: through December 31, 2026, Flash TTS audio output costs $9 per million tokens, and Flash-Lite TTS costs $6; the prior Gemini 3.1 Flash TTS Preview was priced at $20 per million tokens. Starting January 1, 2027, the standard prices for the two models will rise to $18 and $12 per million tokens respectively.

Source

Add to Favorites Download image Share x Copy link

Relevant content

A whale or institutional address transferred 541,000 HYPE tokens to Kraken, generating a profit of $9.52 million after holding the assets for half a month.

According to on-chain data analyst Yu Jin Monitoring, a whale or institutional address transferred 541,000 HYPE tokens (valued at approximately $49.54 million) to Kraken 15 minutes ago. The tokens were withdrawn from Coinbase Prime half a month prior at an average price of $73.9 per unit. Transferred to Kraken at the current $91.5 per token price, the move is estimated to have generated a profit of $9.52 million.

2 minutes ago

Google will send four TPUs into space aboard a SpaceX rocket on October 1.

Beating AI News reports that Google is preparing to launch the first test satellite for its "space AI data center" research project, Project Suncatcher, into low Earth orbit. The satellite will lift off aboard SpaceX’s Transporter-18 mission, marking the first time Tensor Processing Units (TPUs) are tested in a real space environment. Built in partnership with Planet, the satellite is roughly the size of a refrigerator and carries four TPUs. It is not a full-fledged space data center; its main purpose is to test whether the chips can withstand launch vibrations, space radiation, and extreme temperatures. A major challenge in space is heat dissipation, as there is no air for cooling—each TPU can only run continuously for about 15 minutes before needing to cool down. Project Suncatcher’s long-term goal is to deploy a large constellation of TPU-equipped satellites in space, connected via high-speed laser links to form a distributed AI data center. Google plans to launch two additional test satellites in 2027 to verify inter-satellite connectivity.

2 minutes ago

A whale address dormant for four years has woken up and transferred 4,500 BTC.

According to Lookonchain’s monitoring, a whale address bc1qln transferred out 4,500 BTC (approximately $381.38 million) after remaining dormant for over four years.

2 minutes ago

Metalpha withdrew 11,179 ETH from Binance one hour ago.

According to Lookonchain's monitoring, Metalpha withdrew 11,179 ETH (approximately $30.12 million) from Binance an hour ago.

2 minutes ago

Metalpha Withdraws 11,179 $ETH ($30.12M) From Binance