Exclusive Look: Google Gemini 3 Unveiled on AI Studio Before Launch

Google is gearing up to launch Gemini 3, its latest large language model, which is set to debut on AI Studio in the near future. This new model promises enhanced capabilities and control for developers, researchers, and students utilizing Gemini models.

AI Studio provides a platform for building apps with Gemini models, offering users the ability to fine-tune various parameters such as context size and temperature. The current model available on AI Studio is Gemini 2.5 Pro, which already offers advanced control features.

Gemini 3 will initially launch on AI Studio before expanding to gemini.google.com. As the rollout approaches, details about Gemini 3’s capabilities have started to surface, with references to how temperature settings can impact the model’s reasoning abilities.

Prior to its official release, hints of Gemini 3 have been detected on AI Studio, hinting at the importance of maintaining the default temperature setting of 1.0 for optimal performance.

Gemini 3 and Nano Banana 2: Upcoming Models on the Horizon

Google has been actively testing its upcoming models, with Gemini-3-Pro making an appearance on Vertex AI, a cloud platform for AI development. The model, identified as “gemini-3-pro-preview-11-2025,” indicates Google’s commitment to innovation in the AI space.

In addition to Gemini 3, Google is also working on Nano Banana 2, codenamed “GEMPIX2,” which is expected to excel in generating AI images. The release of Nano Banana 2 is anticipated by December 2025, further expanding Google’s AI model offerings.

Stay tuned for more updates from Google as the company prepares to unveil these cutting-edge models in the near future.