Gemini 4 is in post-training, and Google aims to ship it before year end
This story was updated after publication (Oct. 9, 2026). Details
Koray Kavukcuoglu, the new senior vice president of Google DeepMind, said at The Information’s summit that the model has moved into early post-training. Pre-training began on July 21, so reaching post-training in about two months is a compressed timeline. The focus is coding, autonomous agents and long-horizon agentic workflows. No firm date was given; the target is well before the end of the year.
Sources
- The Information, “Google Nears Release of Flagship Gemini 4 AI Model”, (theinformation.com)
- Logan Kilpatrick (X), post on X announcing the start of Gemini 4 pre-training, (x.com)
- Google, “The next chapter of our AI momentum”, (blog.google)
- Google, “Gemini 4 Argon: our next era of frontier intelligence”, (blog.google)
Corrections and updates
- After this story was published, Google announced the model as Gemini 4 Argon on September 30. It is rolling out first to a set of trusted cyber defenders through Google’s Fairwind Program, with no date yet for developers, enterprises and consumers.
About this story
This story was posted on Instagram by @jarrus.tech on Sept. 26, 2026.
Spotted an error in this story? [email protected] · Instagram
Short link: thejarrus.com/en/h8-29
This story in Turkish: Gemini 4 son eğitim aşamasında, Google yıl sonundan önce çıkarmayı hedefliyor
Related
Google introduces Gemini 4 Argon, the first model in its Gemini 4 series
Google introduced Gemini 4 Argon on September 30. It can write up to 1 million tokens per response and is built for coding, enterprise work and cybersecurity.