Google AI’s Gemini 4 model is in the early days of post‑training and Google plans to release it “much earlier” than the end of 2026, Computerworld reported on Sep 25, 2026.
Computerworld quoted Google DeepMind head Koray Kavukcuoglu saying Gemini 4 is in post‑training, the stage where a base model is refined to behave reliably, and that it should be released “much earlier” than the end of this year.
The report noted that Gemini 3.5 Pro remains delayed while Google has been updating smaller, lower‑cost Gemini Flash models aimed at routine tasks; Computerworld framed Gemini 4 as a high‑end release that could shift that balance.
Computerworld also said Kavukcuoglu recently replaced DeepMind founder Demis Hassabis as head of the Google business unit, a leadership change mentioned in the same article as context for the timing of Gemini 4.
Analysts and observers have speculated the model could arrive as early as October, the piece reported, but Computerworld attributed the release timeline specifically to Kavukcuoglu’s comment at an industry summit rather than to a firm Google product announcement.
Computerworld contrasted Google’s cadence with competitors, saying OpenAI has issued several frontier model updates this year while Google has focused more public updates on smaller models; the article cited recent Gemini Flash launches from July as evidence Google kept high‑end updates quieter.
What this means for businesses that tie products or stock expectations to Google AI depends on Google’s formal announcement and rollout schedule; Computerworld’s story presents Kavukcuoglu’s timeline as the clearest public signal so far but stops short of a confirmed launch date from Google.
Sources
- Computerworld — reported Koray Kavukcuoglu saying Gemini 4 is in post‑training and should arrive “much earlier” than year‑end, and provided context about Gemini 3.5 Pro delays and July Gemini Flash releases (Sep 25, 2026).
