Google unveils Gemini 3, its most powerful AI model yet, with major gains in reasoning and coding capabilities

Google unveils Gemini 3, its most powerful AI model yet, with major gains in reasoning and coding capabilities

In a significant advancement in artificial intelligence, Google has introduced Gemini 3, its latest foundation model, now available via the Gemini app and AI search features. This launch comes just seven months after the release of Gemini 2.5, solidifying Google's position as a frontrunner in the competitive AI landscape. The timing of Gemini 3's unveiling is particularly noteworthy, coinciding closely with OpenAI's rollout of GPT 5.1 and Anthropic's Sonnet 4.5, highlighting the rapid pace of development in state-of-the-art AI models. Upcoming is a more specialized version called Gemini 3 Deepthink, which will be offered to subscribers of Google AI Ultra after passing additional safety checks in the coming weeks. According to Tulsee Doshi, Google’s head of product for the Gemini model, Gemini 3 represents a remarkable leap in reasoning capabilities. "It’s responding with a level of depth and nuance that we haven’t seen before," Doshi stated. Early benchmark results appear to back this assertion, as Gemini 3 has achieved a score of 37.4 on the Humanity’s Last Exam benchmark, surpassing the previous record of 31.64 held by GPT 5 Pro. Moreover, Gemini 3 has also topped the LMArena leaderboard, which is based on user satisfaction. Google reports that the Gemini app currently boasts over 650 million monthly active users, with 13 million developers incorporating the model into their projects. This underscores the integral role that Gemini plays within Google's expansive AI ecosystem. In conjunction with the main model's launch, Google has unveiled Antigravity, a novel coding interface powered by Gemini. This tool merges a chat-style prompt with a terminal and browser view, enabling developers to write, test, and refine applications in real time. DeepMind CTO Koray Kavukcuoglu emphasized the agent's collaborative capabilities, stating, "The agent can work with your editor, across your terminal, across your browser to ensure that it helps you build that application in the best way possible." With Gemini 3 making waves in benchmark rankings and the anticipation of Deepthink, Google is clearly signaling its commitment to leading the charge in the race for advanced AI technologies.

Sources : Business Today

Published On : Nov 19, 2025, 03:50

AI
Nadella Challenges AI Labs on Data Usage and Model Training Ethics

In a recent post, Microsoft’s CEO Satya Nadella subtly criticized AI labs, particularly targeting firms like Anthropic, ...

Business Insider | Jul 13, 2026, 05:00
Nadella Challenges AI Labs on Data Usage and Model Training Ethics
Computing
Meta's Ambitious $50 Billion Data Center Expansion in Louisiana Set to Transform AI Landscape

Meta is significantly ramping up its Hyperion data center initiative in rural Louisiana, with the total investment proje...

CNBC | Jul 13, 2026, 09:45
Meta's Ambitious $50 Billion Data Center Expansion in Louisiana Set to Transform AI Landscape
Mobile
iPhone 18 Pro Max Set for a Price Surge: What Consumers Should Know

Apple is gearing up to unveil its highly anticipated iPhone 18 Pro series in just two months, and speculation about a si...

Business Today | Jul 13, 2026, 07:35
iPhone 18 Pro Max Set for a Price Surge: What Consumers Should Know
Startups
Helsing Secures $1.8 Billion to Enhance AI-Driven Defense Solutions

Helsing, a burgeoning defense startup based in Europe, announced on Monday that it has successfully secured $1.8 billion...

CNBC | Jul 13, 2026, 08:05
Helsing Secures $1.8 Billion to Enhance AI-Driven Defense Solutions
Gadgets
Introducing the Hisense A10: A Game-Changer with Detachable Dual Displays

Hisense has unveiled an innovative smartphone, the Hisense A10, featuring a unique dual-display system that includes a d...

Business Today | Jul 13, 2026, 10:00
Introducing the Hisense A10: A Game-Changer with Detachable Dual Displays
View All News