Google DeepMind is close to releasing its next flagship artificial intelligence model, Gemini 4, according to the division's new leader. Koray Kavukcuoglu said in his first media appearance as head of DeepMind that the model is in its refinement stage. He told The Information that Google aims to launch Gemini 4 well before the end of 2026.
Kavukcuoglu said the company intends to release an early post-training output as soon as possible because the results have the team excited. He added that Google will keep up a fast pace of iterations on the model going forward. The remarks point to a more consistent release schedule for Google's AI products after a change at the top of DeepMind.
The launch would be Google's first new flagship AI model since the Gemini 3 series arrived in November 2025. Since then, OpenAI has released GPT-6 and Anthropic has launched its Mythos series models. Both outperform Gemini 3, giving those rivals a substantial lead over Google in the AI race.
Google CEO Sundar Pichai said at the I/O event in May that a major Gemini 3.5 Pro update would arrive in June to help close that gap. That model was never rolled out. Kavukcuoglu said the company stepped back to focus on faster, less powerful Flash models instead of shipping Gemini 3.5 Pro.
Former AI head Demis Hassabis stepped down in August after saying great progress was being made on Gemini 4. The leadership change came as Google worked to catch up with competing developers on flagship releases. Kavukcuoglu took over the DeepMind division during that period.
When asked whether he worried about Google falling behind its rivals in AI development, Kavukcuoglu told The Information that he considers it a certainty that Google will always be at the frontier. His comments mark the first public outline of the company's plans for Gemini 4 under his leadership. Google has not announced a specific release date for the model.
More AI news from TechManNews.
