Google DeepMind chief Koray Kavukcuoglu said Gemini 4 has entered the early stages of post-training, putting Google closer to releasing its next flagship AI model.. Kavukcuoglu said at The Information’s AI Agenda Live Summit on Wednesday that Gemini 4 is expected to launch “much earlier” than the end of 2026..
He said post-training is the phase in which a base model is refined with human feedback, reinforcement learning and specialized data to improve reasoning, safety and reliability.. “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,” Kavukcuoglu told The Information..
He added that Google would “continue the fast-paced iterations” after that release.. The remarks were Kavukcuoglu’s first public appearance since he took over DeepMind following the August departure of co-founder Demis Hassabis.. Google has not released a new flagship model since the Gemini 3 series in November 2025, while OpenAI launched GPT-6 Astra on Sept.
3 and Anthropic released Claude Fable 5.1 on Sept. 1.. Google Chief Executive Sundar Pichai said at the company’s I/O event in May that a Gemini 3.5 Pro update would arrive in June, but the model was never released..
Kavukcuoglu said Google “took a little bit of a step back” to focus instead on faster, less powerful Flash models.. Alphabet shares fell nearly 5% over two sessions through Wednesday as Meta Platforms gained momentum with its Muse AI agent, which climbed to the top of U.S. free app charts during Meta’s Connect event..
When asked whether he was concerned about Google falling behind, Kavukcuoglu told The Information: “In my mind, it’s a certainty that we are always gonna be at the frontier.”. Google plans to release Gemini 4 in stages, starting with an early version and then iterating based on user feedback rather than waiting for a more polished launch..
Featured image credit. Tags: FeaturedGemini 4Google Deepmind
