Google’s new DeepMind leader, Koray Kavukcuoglu, says the company is refining its next flagship AI model, Gemini 4, for an early release.

Google is nearing the launch of its Gemini 4 AI model, according to Koray Kavukcuoglu, the new leader of the company's DeepMind division. In his first media appearance, Kavukcuoglu stated that the model is in the refinement stage and that Google aims to release it well before the end of the year.

Refinement and Release Strategy

Kavukcuoglu told The Information that Google intends to release an early post-training output as soon as possible, citing excitement over current results. He emphasized that the company plans to maintain fast-paced iterations moving forward. This shift follows a leadership change at DeepMind, where former AI head Demis Hassabis stepped down in August.

Competitive Landscape

Google has not released a new flagship AI model since the Gemini 3 series in November 2025. In the interim, competitors OpenAI and Anthropic have launched the GPT-6 and Mythos series, respectively. Google previously canceled a planned June release of a Gemini 3.5 Pro update, with Kavukcuoglu explaining that the company chose to prioritize the development of faster, less powerful Flash models instead. Despite the competition, Kavukcuoglu stated he is certain Google will remain at the AI frontier.