Google has unveiled Gemini 3, its latest cutting-edge artificial intelligence model, marking the beginning of a widespread release across its range of products. The model is now accessible to the general public through the Gemini app and to subscribers of Google AI Pro and Ultra who are utilizing AI Mode in Search. Developers can leverage the model via the Gemini API in AI Studio, the newly introduced Google Antigravity development platform, and Gemini CLI. Additionally, enterprise customers have the option to utilize the model through Vertex AI and Gemini Enterprise.
Gemini 3 is hailed by Google as its most advanced model yet, boasting enhanced reasoning capabilities and a deeper comprehension of intricate queries. According to the company, Gemini 3 surpasses its predecessor, Gemini 2.5 Pro, in standard benchmarks for coding, mathematics, and multimodal tasks.
In addition to the primary model, Google is gearing up to launch an upgraded version named Gemini 3 Deep Think, which will undergo rigorous safety assessments before being accessible to Google AI Ultra subscribers. Google has underscored that Gemini 3 has undergone thorough safety evaluations to ensure its reliability and security.