Google launches Gemini 3, says new AI model grasps nuance, context better

Google has unveiled Gemini 3, its newest AI model designed to better understand nuance, context and intent, with improved safety features and a new Deep Think reasoning model to boost performance

Google on Tuesday announced the release of Gemini 3, the company’s latest update to the flagship artificial intelligence (AI) model, which will also be integrated in the AI mode of the Google search engine, the company said in a blog post.

 

“The Gemini App surpasses 650 million users per month, more than 70% of our cloud customers use our AI, 13 million developers have built with our generative models, and that is just a snippet of the impact we are seeing,” Google and Alphabet Chief Executive Officer Sundar Pichai said in the post.

 

The Gemini 3, Google said, is built to grasp nuance and depth, better understands the context and intent behind users’ requests, and “peeling apart the overlapping layers of a difficult problem”.

 

The responses in Gemini 3 Pro will be “smart, concise, and direct, trading cliche and flattery for genuine insight”. 

 

“It acts as a true thought partner that gives you new ways to understand information and express yourself, from translating dense scientific concepts by generating code for high-fidelity visualisations to creative brainstorming,” Google said in the blog post.

 

Gemini 3, Google said, also has enhanced safety features, such as increased resistance to prompt injections and improved protection against misuse via cyberattacks.

 

Along with the launch of Gemini 3, Google on Tuesday also said it would introduce Gemini 3 Deep Think, an enhanced reasoning model to push the performance of its flagship AI models.

 

The launch of Google’s flagship AI models comes just days after its rival, OpenAI, unveiled an upgraded version of its own flagship AI model and chatbot, ChatGPT.

 

The latest model, ChatGPT 5.1, would be “warmer, more intelligent, and better at following,” the company had then said in a blog post.

 

“What you will notice is that answers across GPT-5.1 feel both smarter and more natural in tone. We plan to roll it out gradually over the next few days to help keep performance stable for everyone,” OpenAI had then said.