Google DeepMind Chief Says Gemini 4 Enters Final Development Stage
Google DeepMind's new leader, Koray Kavukcuoglu, revealed at The Information's AI Agenda Live Summit that the Gemini 4 model is in the early post-training phase, signaling an imminent release. This comes as Google seeks to catch up with rivals Anthropic and OpenAI in the AI model race. Kavukcuoglu made the remarks in his first media appearance since taking the helm.
Koruk Kavukcuoglu, Google DeepMind’s newly appointed chief, used his first public appearance since taking the role to signal that Gemini 4 is nearing completion. Speaking at The Information’s AI Agenda Live Summit, he described the model as being in its early post-training phase—a stage that typically precedes final testing and deployment. The update suggests a launch could come soon.
The announcement arrives as Google works to close a perceived gap with rivals Anthropic and OpenAI, both of which have released high-profile models in recent months. Kavukcuoglu’s remarks offer a rare glimpse into the company’s internal timeline, though no specific release date was provided.
The imminent arrival of Gemini 4 could reshape competitive dynamics in the AI sector, potentially accelerating adoption across industries that rely on advanced language models. Businesses and developers may gain access to improved capabilities, while consumers could see more capable assistants and tools. However, rapid deployment also raises questions about safety, reliability, and the pressure on smaller players to keep pace. The impact will likely depend on how the model performs in real-world use and how rivals respond.