AI Product Updates: New Features and Releases

Google DeepMind announced two new text-to-speech models, Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS.

Grok confirmed that Grok 4.7 is now available in Warp and the Warp Agent CLI, letting users connect a Grok subscription and run agents in the terminal.

Warp announced that Grok 4.7 has landed in Warp and the Warp Agent CLI, and users can connect a Grok subscription to drive agents with the model in the terminal.

Belong added a step-by-step Telegram setup guide to its dashboard and fixed Sui transaction decoding, NFT collection creation, wallet login and other issues.

Google released two new text-to-speech models with custom voice design, dual-speaker screenplay control and line-by-line delivery, now available in AI Studio and the API.

Flash TTS focuses on custom accents and voice design, while Flash-Lite TTS targets high-throughput use and can reuse created styles or the official voice library.

The new TTS models let users adjust pacing, emotion and cues like laughs or pauses line by line, and all generated audio carries a SynthID watermark for identification.

Google calls it its most expressive audio generation model yet, available to try in the Gemini API and AI Studio.

The new models launched the same day in AI Studio, Gemini Notebook and the Gemini API, with Google Vids and Gemini Enterprise to follow.

The models support custom voices in 100+ languages, two-speaker dialogue, nonverbal cues and voice cloning from a 30-second sample, with SynthID watermarks on all audio.

Flash TTS ranks first on Hume AI's voice design and overall quality leaderboards with Flash-Lite second, and shows gains in long-form stability and two-speaker dialogue.

Google DeepMind released Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS, two audio generation models focused on expressive speech.