Google launches Gemini 3 as latest move in AI race
by Jake Thomas · UPINov. 18 (UPI) -- Google announced Tuesday it was rolling out the latest iteration of its Gemini AI model as the search engine giant tries to keep up with competitor OpenAI.
Sundar Pichai, CEO of Google and parent company Alphabet, made the announcement in a blog post touting Gemini 3 as an AI model "built to grasp depth and nuance" that will help users get what they need "with less prompting."
He wrote that it could be used to translate recipes handwritten in different languages into a shareable family cookbook or develop a training program after watching a user's pickleball matches.
Gemini 3 is now built into the Gemini app and Google's search products, as well as products geared toward developers, including AI Studio and AI Vertex.
Related
- Google's experimental Mixboard uses AI to visualize ideas
- Google announces AI-heavy Pixel 10 models, plus its Pixel Watch 4
- Google relaunching problematic tool for AI-generated images
The Gemini app now has more than 650 million monthly users and Google AI Overviews has 2 billion monthly users, according to the company.
The release comes months after Google made the chatbot free to all users and unveiled Gemini 2.5.
Demis Hassabis and Koray Kavukcuoglu, two of Google's AI executives, wrote in another blog post that Gemini 3 Pro would trade "cliche and flattery for genuine insight," a thinly veiled reference to criticism that ChatGPT is overly sycophantic to users.
Gemini 3's release follows the debut of OpenAI's Atlas web browser in October, which is designed to incorporate the company's chatbot into daily Internet use and could be a competitor to Google.
OpenAI launched GPT-5, the most advanced version of ChatGPT, in August, the same month the company announced it would pass the 700 million weekly user mark.