Google has launched Gemini 3, the most recent model of its massive language mannequin (LLM), now out there via the Gemini app and Google’s AI-powered search.
In keeping with a weblog publish revealed on November 18, Gemini 3 is described as Google’s most succesful mannequin up to now.
A research-focused model referred to as Gemini 3 Deepthink can even be launched quickly for Google AI Extremely customers, as soon as it finishes further security evaluations.
Do you know?
Subscribe – We publish new crypto explainer movies each week!
10 Greatest Crypto Scams & Learn how to Keep away from Them (ANIMATED)
Tulsee Doshi, Google’s head of product for the Gemini mannequin, stated in a weblog publish:
With Gemini 3, we’re seeing this huge bounce in reasoning. It’s responding with a stage of depth and nuance that we haven’t seen earlier than.
Early checks present that Gemini 3 scored 37.4 on the Humanity’s Final Examination benchmark. The earlier finest rating, held by GPT-5 Professional, was 31.64. It additionally led the LMArena benchmark, which measures consumer satisfaction with AI responses.
The Gemini app at present serves over 650 million customers every month, in line with Google, and about 13 million builders use its instruments of their each day work.
Moreover, Google launched a brand new coding device referred to as Google Antigravity. The system combines options of a chat assistant, command line, and internet browser. It permits programmers to put in writing and take a look at code in a single place.
Google Antigravity helps a method of “agentic coding”, just like newer AI-based improvement instruments like Warp and Cursor 2.0.
Google not too long ago launched its Gemini AI assistant to Maps in India. What options does it supply? Learn the complete story.
