In November 2025, Gemini 3 was launched by Google, marking what the company describes as the next major leap in artificial intelligence capabilities.

Gemini 3 brings together state-of-the-art reasoning, multimodal understanding and agentic capabilities, enabling users to interact with AI in far more powerful ways than before. According to Google, Gemini 3 Pro — the entry‑level release of the new version — already surpasses previous models across benchmarks for reasoning, coding, factual accuracy and multimodal tasks.

A key feature of this release is the introduction of Nano Banana Pro for image generation and editing, coupled with Gemini 3’s reasoning engine. This lets users transform prompts, data or even sketches into complex infographics, designs or visual outputs — giving AI a far more creative and practical edge.

Developers are also being empowered with new tools: alongside Gemini 3, Google released Google Antigravity — an “agent‑first” coding platform that allows the AI to function as a coding assistant with increased autonomy, orchestrating tasks, generating real‑world codebases, and handling complex workflows.

This release comes at a time when AI competition is heating up. The surge in capability — especially in reasoning, code generation, and multimodal tasks — threatens to upend previous assumptions about what AI can reasonably do for users, engineers, creators, and businesses. With Gemini 3, Google is reshaping expectations, showing that the next generation of AI isn’t just about chatter — it’s about real, versatile intelligence that spans text, code, images and more.