In the past week, Google offered several hints about its next major model, Gemini 4, after the planned Gemini 3.5 Pro release missed a June launch. The company has been advancing multiple Gemini 3.x variants while preparing a more ambitious next-generation model.
Gemini 4
The first public mention of the unsurprisingly named “Gemini 4” came alongside the rollout of new Gemini 3.x Flash variants, including Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber. While Google continues to test Gemini 3.5 Pro with partners and plans to release it “as soon as it’s ready,” work is already underway on the next generation of models. Google describes this effort as its “most ambitious pre-training run yet,” signaling a substantial increase in scale and ambition compared with prior releases.
During Alphabet’s Q2 2026 earnings call, CEO Sundar Pichai emphasized the company’s ongoing commitment to advancing frontier AI capabilities. He noted that the frontier of AI is a rapidly evolving space and that Google has maintained frontier-level models across several dimensions, even while acknowledging areas that need improvement.
“Look, I think the frontier is an incredibly dynamic space, and it’s fiercely moving forward. At any given moment when you take a snapshot, it feels dynamic. We have had clearly frontier models. There are many attributes on which we are still at the frontier. There are areas where we have acknowledged we need to improve. Coding and agentic coding is an example of that, and I think the teams are very, very focused on it.”
Pichai went on to affirm Google’s dedication to staying at the cutting edge: “We are both very committed and very confident of being at the frontier.” He added that the company aims to match the frontier as it will be when Gemini 4 launches, and that considerable compute and engineering effort are being applied toward that goal.
In response to questions about balancing internal and external demand for Tensor Processing Units (TPUs), Pichai explained that prioritizing TPU allocation for frontier-level AGI development is foundational to Google’s strategy. Ensuring adequate compute for developing these advanced models is being treated as a top priority.
“In terms of allocating TPUs, our first priority is making sure we are allocating what we need to compete at the frontier in terms of AGI development. And so that is the foundation for everything we do.”
Perhaps the most concrete detail shared was that Gemini 4 will rely on a much larger base model than prior releases. Pichai described Gemini 4 as an ambitious training effort currently underway and expressed confidence in the team’s progress and in the reception the model will receive once it becomes available externally.
“For the next generation of frontier, you’re going to need much larger base models. We are now training Gemini 4, and we’re being very ambitious with it. I am very excited by the progress I’m seeing internally on Gemini 4, and I’m confident that people will be pleased when we are putting it outside.”
Based on Google’s historical release cadence, many observers expect Gemini 4 to debut toward the end of the year, potentially in November or December. That timing aligns with major model releases and the company’s pattern of deploying substantial updates before year-end.
Gemini Flash
Alongside discussion of Gemini 4, Pichai also signaled continued development on the Gemini 3.x Flash series. These Flash models aim to iterate quickly on capabilities such as agentic coding and other targeted improvements. Google indicated a goal of releasing incremental model updates at a fast pace, potentially approaching a near-monthly cadence to push performance forward and respond to real-world feedback.
In short, Google is balancing short-term iterative improvements through the Gemini 3.x Flash line with a long-term, large-scale effort to train Gemini 4. The company’s focus on increased TPU allocation and larger base models underscores its objective to remain competitive at the leading edge of AI and AGI research.