Skip to main content

Google's Gemini 4 Is Almost Here—And It's Coming Sooner Than You Think

Google's Next Big Thing: Gemini 4 Enters Home Stretch

At The Information's AI Agenda Live event on September 23, Demis Hassabis, the head of Google DeepMind, dropped a pretty big hint: Gemini 4 is coming, and it's already in "early post-training." If you're not familiar with the term, post-training is the stage where a base model gets fine-tuned and its behavior becomes more reliable. In other words, the heavy lifting is done—now they're just polishing it up.

Hassabis didn't hold back his excitement. He said he hopes Gemini 4 will launch "well before the end of the year" and that the team is already seeing promising early results. "Because we saw the results and were very excited," he explained. After the initial release, Google plans to keep iterating quickly.

Skipping 3.5 Pro, Going All-In on 4

So what happened to Gemini 3.5 Pro? Back in May, Google CEO Sundar Pichai promised a major update in June. That didn't happen. Hassabis clarified that after Gemini 3 and 3.1, the team decided to "take a step back" and focus on the smaller, cheaper Flash models instead. "For us, the most important thing at that time was to accelerate learning as much as possible," he said. He didn't say 3.5 Pro was canceled, but he made it clear: Gemini 4 is now the priority.

This will be Google's first next-gen flagship since the Gemini 3 series last year. In February, they rolled out an upgraded 3.1, followed by several Flash models that were lighter on cost but also less complex in reasoning.

Already Testing It Internally

Gemini 4 isn't just a lab experiment. Google engineers have already used it to run the Antigravity AI programming tool internally. Right now, the model is in the safety testing and protection mechanism phase—standard procedure before any big launch.

When asked if he's worried about falling behind competitors, Hassabis didn't flinch. "100% confidence in the team," he said. "Our team is incredibly strong, and I have the greatest trust in them; we will always be at the forefront of technology."

The AGI Question? Hassabis Says We're Asking the Wrong Thing

Unlike some rival executives, Hassabis didn't take a clear stance on whether the industry has achieved AGI (artificial general intelligence). Instead, he pushed back on the question itself: "Discussing whether AGI has been achieved is not the right direction; the real issue is whether we can build trustworthy agents." That's a subtle but important shift—focusing on practical, reliable AI rather than philosophical labels.

With Gemini 4 in post-training and internal testing underway, all signs point to a launch that could shake up the AI race. And if Hassabis's excitement is any indication, Google might just have something special on its hands.

Key Points:

  • Gemini 4 is in early post-training and could launch well before the end of the year.
  • Google is skipping Gemini 3.5 Pro to focus on Gemini 4.
  • The model is already being used internally for coding tools and safety testing.
  • Hassabis expresses 100% confidence in the team and shifts the AGI debate toward building trustworthy agents.