Google launches its latest AI model, Gemini 3

Google has launched Gemini 3, its newest and most advanced artificial intelligence model, initiating a broad rollout across its product ecosystem.

The model is now available to general users in the Gemini app and to Google AI Pro and Ultra subscribers using AI Mode in Search. For developers, access is provided through the Gemini API in AI Studio, the new Google Antigravity agentic development platform, and Gemini CLI. Enterprise customers can use the model through Vertex AI and Gemini Enterprise.

Google describes Gemini 3 as its most capable model to date, featuring improved reasoning abilities and a better understanding of complex requests. The company states it outperforms the previous Gemini 2.5 Pro version across standard benchmarks for coding, mathematics, and multimodal tasks.

Alongside the main model, Google is preparing an enhanced version called Gemini 3 Deep Think, which will undergo additional safety evaluation before becoming available to Google AI Ultra subscribers. The company also emphasised that Gemini 3 has undergone comprehensive safety testing.