Google Gemini
Google is moving closer to rolling out its next premier artificial intelligence (AI) model, Gemini 4, the newly appointed head of its DeepMind research arm has revealed. Kavukcuoglu revealed that Gemini 4 has formally entered the initial stages of post-training, an essential engineering phase where raw base models are fine-tuned to ensure dependable, consistent behaviour. The update on Gemini 4 arrives as Google rolls out two advanced text-to-speech (TTS) systems: Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS. Gemini 3.8 Flash-Lite TTS Developed for high-volume, low-cost deployments. The new tools build on Google’s wider Gemini Audio portfolio that includes existing specific features like 3.5 Live Translate, 3.5 Transcribe, 3.8 Live, and 3.8 Live Extended Thinking.
the executive noted that he hopes to release the flagship model “much earlier” than the close of the year While testing is ongoing. The releases are built to support developers, content creators and businesses, while bringing upgraded audio capabilities to Google products such as Google Vids and Gemini Notebook, according to the company. The upcoming release will mark a critical milestone as the company works to close the gap with industry rivals Anthropic and OpenAI. Speaking Wednesday at The Information’s AI Agenda Live Summit in his first public media appearance as top leader, Koray Kavukcuoglu, senior vice president at Google DeepMind and Google’s chief AI Architect, shared key updates on the model’s development schedule. Described as the company’s most expressive speech-generation tools to date, the dual models shift voice production away from rigid pre-recorded presets into an adaptive creative environment. It’s meant for large scale dubbing pipelines, automated audio production, and conversational voice bots, with fine-grained control over delivery speed, vocal tone, and emotional inflection.