Google is preparing to release Gemini 3.5 Pro in June, according to a report that says the company’s latest flagship AI model is nearing general availability after its debut at Google I/O in May.
The model is positioned as Google’s top-tier offering for demanding AI work, with a reported 2-million-token context window, a Deep Think reasoning mode, and support for multimodal tasks across text, images, and other formats. Google has not yet broadly shipped the model, but it is expected to reach consumer subscribers first through the company’s Pro and Ultra plans.
Gemini 3.5 Pro appears intended to handle the most challenging workloads in Google’s AI lineup. The reported 2-million-token context window would allow the model to process very large amounts of information in one session, such as lengthy documents, code repositories, or extended back-and-forth conversations. In practical terms, that kind of capacity could make the model more useful for enterprise users and developers dealing with large datasets or complicated workflows.
Deep Think is another feature that points to Google’s emphasis on higher-end reasoning. Rather than focusing only on quick responses, the mode suggests a model designed to spend more effort on difficult problems. The source material indicates that this capability will be available to Ultra subscribers when the model rolls out.
Google has also folded responsibilities that were previously assigned to its Ultra tier into Gemini 3.5 Pro. That includes the hardest reasoning tasks and more advanced multimodal use cases, making Pro the central flagship in the company’s model family.
Although Google has signaled a June launch, the model was still not generally available in early June. At Google I/O, CEO Sundar Pichai told the audience to wait about another month, according to the report. That left attendees disappointed, but it also underscored that the company is still controlling the rollout closely.
The model has reportedly been used internally and made available in limited preview through Google’s enterprise platform. Even so, the source material notes that public claims about Gemini 3.5 Pro’s capabilities remain unverified until independent testers can examine the model after release. Launch schedules in the AI industry can also slip, so June should be viewed as the target rather than a firm guarantee.
Gemini 3.5 Pro follows Gemini 3.5 Flash, a faster and cheaper model that Google shipped earlier in the spring. Flash reportedly improved on the previous generation’s Pro model in coding and agent-like tasks, but it lagged on the most difficult reasoning challenges. Pro is meant to close that gap by focusing on capability instead of speed or low cost.
The pricing structure is expected to mirror earlier generations, with Pro costing roughly 10 times more than Flash. That would place it at about $15 per million input tokens and $60 per million output tokens, based on the report. Google is expected to offer the model first to users of its $20-a-month Pro subscription and its $250-a-month Ultra plan.
The launch comes as major AI developers continue to push more capable models into the market, while pricing pressure increases from rivals across the U.S. and China. Google’s strategy has been to cover multiple tiers, from lower-cost models to premium systems aimed at the most demanding users.
Gemini 3.5 Pro is meant to be the centerpiece of that strategy. If it performs as described, it could give Google a stronger position in frontier AI reasoning and multimodal work. If the rollout is delayed or the model falls short of expectations, it could reinforce the perception that Google announces ahead of delivery.
For now, Gemini 3.5 Pro remains one of the most closely watched AI releases on Google’s calendar. The key questions are straightforward: when will it ship, and will the promised context length and reasoning mode translate into measurable real-world gains once the model is in users’ hands?