Google DeepMinds new chief said Gemini 4 is already in early post-training and that the company wants to ship an early version well before the end of 2026, The Verge reported on Thursday, September 24, 2026, citing Koray Kavukcuoglu’s first media appearance as DeepMind leader in an interview with The Information.
Kavukcuoglu told The Information that Gemini 4 is in its refinement stage and that Google aims to launch “much earlier” than year-end. Our intention is to, like — as soon as possible — to release an early post-training output because we see the results and we are excited,” he said, adding that Google would “continue the fast-paced iterations” after that first cut.
The comments land after a leadership change at DeepMind. Demis Hassabis stepped down in August after saying “great progress was being made on Gemini 4. Google has not shipped a new flagship model since the Gemini 3 series in November 2025. In that window, OpenAI’s GPT-6 and Anthropic’s Mythos-class models arrived and, per Verges reporting, outperform Gemini 3 on the current frontier race.
CEO Sundar Pichai told Google I/O in May that a major Gemini 3.5 Pro update would arrive in June to help close the gap, but that model never rolled out. Kavukcuoglu said the company “took a little bit of a step back” to prioritize faster Flash models instead of 3.5 Pro. Asked whether he worried Google was falling behind, he told The Information: “In my mind, it’s a certainty that we are always gonna be at the frontier.”
This brief covers Kavukcuoglu’s Information interview as reported by The Verge — post-training status, the early-release intent, the Flash detour, and the frontier claim. It does not invent a firm launch month, claim Gemini 4 is live, or restate AI Tech Daily’s separate Gemini 3.8 Flash TTS coverage.