Google Launches Gemini 3 as Rivalry With OpenAI Deepens

Google has taken another major leap in its AI journey with the launch of Gemini 3, a new model designed to power its growing family of AI tools. The company began rolling out the upgrade on Tuesday to a small group of subscribers, with a wider release expected soon. It arrives less than a year after Gemini 2.0 and only eight months after Gemini 2.5, signalling how rapidly the AI race is shifting.

Sundar Pichai said the new release marks a moment where AI is no longer just parsing text or images but learning to “read the room,” as he put it in a post on Tuesday. “Starting today, we’re shipping Gemini at the scale of Google,” he wrote. That claim reflects Google’s expanding AI footprint, as the company now reports 650 million monthly users on the Gemini app and 2 billion users engaging with AI Overviews.

A New Chapter in Google’s AI Competition

The arrival of Gemini 3 comes at a time when Google and OpenAI are locked in an intense rivalry. OpenAI introduced GPT-5 in August and followed it with two updates last week, one said to be “warmer, more intelligent,” and another aimed at speed and stability. Although both companies continue to refine their models, the pressure to stay ahead is pushing them to release improvements at a faster pace than ever.

Google claims that Gemini 3 offers a deeper understanding and stronger context handling. According to Pichai, the model is “built to grasp depth and nuance”, allowing users to get clearer answers with less prompting. Demis Hassabis, chief executive at DeepMind, said the system now focuses on “genuine insight” rather than the flattery and clichés that critics say plague many modern chatbots.

The company explained that while Gemini 3 will take on more complex tasks, older models will still handle smaller or routine queries. The shift reflects how Google is layering its systems to meet different user needs.

Alphabet and other tech giants are also pouring heavy investment into AI infrastructure. In their latest earnings, Google, Meta, Microsoft and Amazon all raised their capital expenditure forecasts, with combined spending expected to exceed $380 billion this year.

Gemini 3 Brings New Tools for Users and Developers

Google is not only targeting everyday users with Gemini 3, but also developers and enterprise clients. It introduced a new agent platform called “Google Antigravity,” which lets developers produce code at a higher, task-driven level. Josh Woodward of Google Labs described Gemini 3 as the company’s “best vibe coding model ever”, referring to the trend of generating software through conversational prompts.

The model can also create “generative interfaces,” producing results that resemble colourful digital magazines. Google showed an example where Gemini 3 explained the Van Gogh Gallery by adding images, notes and historical details for each artwork. It will also create visual layouts in AI Mode, including tables, grids, and interactive calculators.

Inside the Google ecosystem, developers will access the model through the Gemini API, while corporate teams can use it via Vertex AI. Google says Gemini 3 can streamline onboarding, improve video analysis, and help with factory operations or procurement tasks.

Although the rollout is still early, the company expects adoption to grow quickly as more users test the model across the Gemini app, AI search tools and enterprise products. The question now is whether Google’s newest upgrade gives it enough of an edge in a competition that only gets faster each month.