Google DeepMind is moving ahead with its next major artificial intelligence model, Gemini 4. DeepMind chief Koray Kavukcuoglu has revealed that the model has entered the early stages of post-training, suggesting that Google is preparing for a launch earlier than the end of 2026.
Gemini 4 Development and Launch Timeline
Koray Kavukcuoglu, senior vice president at Google DeepMind and Google’s chief AI Architect, shared the update during The Information’s AI Agenda Live Summit. It was his first public media appearance since becoming the head of Google’s DeepMind research division.
Kavukcuoglu said Gemini 4 has formally entered the initial stages of post-training. This stage is important because engineers fine-tune the base AI model to improve its reliability, consistency and overall performance.
While Google is still testing Gemini 4, Kavukcuoglu said he hopes the company can release the flagship model “much earlier” than the close of the year.
The launch will be closely watched as Google competes with major AI companies including OpenAI and Anthropic. Gemini 4 is expected to become an important part of Google’s next generation of AI products and services.
Google Launches New Gemini TTS Models
The Gemini 4 update comes as Google also introduces two new text-to-speech models, Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS.
Google describes the models as its most expressive speech-generation tools so far. They are designed to give developers, businesses and content creators greater control over AI-generated voices.
Gemini 3.8 Flash-Lite TTS is designed for high-volume and lower-cost uses. Google says it can support large-scale dubbing, automated audio production and conversational voice bots. Developers can also control factors such as speaking speed, vocal tone and emotional expression.
The new models add to Google’s broader Gemini Audio portfolio, which includes features such as 3.5 Live Translate, 3.5 Transcribe, 3.8 Live and 3.8 Live Extended Thinking.
Google is preparing for the next major stage of its AI strategy with Gemini 4. The model has entered post-training, and DeepMind chief Koray Kavukcuoglu hopes for a launch much earlier than the end of 2026. At the same time, Google’s new Gemini TTS models expand its AI audio capabilities for developers and businesses.
