The AI world might be getting two major updates from Google before the year ends. A new leak has revealed that Google Gemini 3 Pro and Nano Banana 2 could both debut in the coming months — bringing big improvements in AI coding, reasoning, and image generation.
As spotted on X (formerly Twitter), the model “gemini-3-pro-preview-11-2025” appeared on Vertex AI, Google’s cloud platform for AI development and deployment. That model name alone aligns with recent reports claiming Gemini 3 Pro will ship in November 2025.
While some users online are claiming early access to the model, it’s highly unlikely. Google typically tests its AI models under strict NDAs, with limited access to internal teams or select enterprise partners.
Google Gemini 3 Pro: A Leap Forward for Coding and AI Agents
The upcoming Google Gemini 3 Pro is said to be optimized for both coding and general-purpose AI tasks, positioning it as a direct competitor to OpenAI’s GPT-5.1 and Anthropic’s Claude Sonnet 4.5.
According to early leaks, Gemini 3 Pro could feature a 1 million token context window, allowing developers to handle massive projects, longer codebases, and more complex reasoning tasks seamlessly.
Currently, Gemini 2.5 Pro is Google’s flagship model — and it’s still one of the strongest performers in the AI space. On the SWE-Bench Verified benchmark for agentic code evaluation, Gemini 2.5 Pro scores 63.8%, while Claude Sonnet 4.5 sits higher at 77%.
If these numbers are any indication, the Google Gemini 3 Pro could close that gap significantly, bringing Google back to the forefront of AI coding performance.
Nano Banana 2: Google’s Next Big AI Image Model
Alongside Gemini 3, Google is quietly developing Nano Banana 2, a new AI model focused on generating highly realistic images.
Codenamed “GEMPIX2,” the model was recently discovered on Google’s Gemini website, sparking speculation that Nano Banana 2 could roll out as soon as December 2025.
The first version of Nano Banana gained attention for its efficiency in generating detailed visuals with minimal processing power. If Nano Banana 2 lives up to the leaks, it could rival OpenAI’s DALL·E 4 and Midjourney v7, combining speed, realism, and accessibility.
It’s another sign that Google wants to dominate both text-based AI and multimodal creativity tools within the same Gemini ecosystem.
Why These Leaks Matter
The Google Gemini 3 Pro and Nano Banana 2 leak matters for several reasons.
First, it highlights Google’s increasingly aggressive timeline for AI innovation — just eight months after the release of Gemini 2.5 Pro. Second, it signals Google’s strategy to build specialized AI systems rather than one all-purpose model.
- Gemini 3 Pro → Focused on coding, productivity, and reasoning tasks.
- Nano Banana 2 → Designed for generating lifelike images and creative visuals.
This division of roles allows Google to create more efficient, purpose-built AI systems that can better compete across different verticals — from business automation to creative media.
Competition Heats Up in the AI Space
The launch of Google Gemini 3 Pro and Nano Banana 2 comes at a time when the AI race is more competitive than ever.
OpenAI is gearing up for its own release of GPT-5.1, reportedly with agentic upgrades and a new Codex terminal for developers. Meanwhile, Anthropic, Mistral, and xAI are all rapidly advancing their models, each targeting unique use cases — from reasoning agents to open-source frameworks.
With Google entering the next phase of its Gemini roadmap, 2026 could shape up to be one of the most defining years for generative AI.
What’s Next for Google Gemini 3 Pro and Nano Banana 2
If the leaks are accurate, we could see Google Gemini 3 Pro debut in November 2025, followed shortly by Nano Banana 2 in December.
Given Google’s pattern, both models are likely to roll out first to enterprise customers and developers through Vertex AI and Gemini Advanced, before expanding to broader user bases.
For now, all eyes are on Google’s next big AI event — where we might finally get the official confirmation.
Key Takeaway
The leaked details around Google Gemini 3 Pro and Nano Banana 2 suggest Google is preparing to make a major comeback in the AI race.
With Gemini 3’s coding prowess and Nano Banana 2’s imaging capabilities, the company is doubling down on its vision of a multimodal AI ecosystem — one that can think, code, and create, all in one platform.
If these models perform as expected, Google might just reclaim its place as the global AI leader heading into 2026.
Don’t miss out on our latest news—follow us for the latest AI news, breakthroughs, and insights that matter.