Google said the information produced by Gemini 3 was 72 percent accurate. (Credit: New York Times)
Google has unveiled Gemini 3, its most advanced AI model yet, marking another push to reclaim dominance in a sector where it once seemed to lag behind. The release comes at a moment when competition with OpenAI and Anthropic has intensified, and expectations from users and investors continue to climb.
A More Capable, More Versatile Model
Gemini 3 promises stronger performance across key tasks. It writes software more accurately, sorts email with greater efficiency, and processes dense business documents with improved clarity. The model also blends text and visuals when dealing with queries related to travel, art, or historical information.
The model will be available on the Gemini app and will power Google’s AI Mode in Search—its conversational search feature launched earlier this year.
Rivals Watch Closely
Inside OpenAI and Anthropic, Google’s newest model has sparked concern. Employees at both companies worry that stronger performance in autonomous coding or image generation could tip the balance of the AI race, according to people familiar with internal conversations.
Analysts agree. Mike O’Rourke, chief market strategist at JonesTrading, said Google’s scale gives it a natural edge. If Gemini wins over developers, he noted, it could affect the long-term prospects of rivals that depend on massive cloud computing deals, including Microsoft and Oracle.
A Costly Race With Uncertain Rewards
Despite rapid progress, the AI boom faces a financial question: will revenue catch up with soaring investment? Most AI systems today help with search queries or automate programming. Yet they rely on huge, power-hungry data centers. McKinsey estimates the industry may spend nearly $7 trillion building such infrastructure by 2030.
Investors are growing cautious. Ben Bajarin, principal analyst at Creative Strategies, said meaningful revenue will come only when AI delivers reliable, high-value use cases. For now, he believes that moment has not arrived.
Accuracy Improves — But Expectations Remain High
Google says Gemini 3 reached 72% accuracy on a standard benchmark test—high for a model of its complexity. The model also outperformed several leading competitors on other evaluations.
Koray Kavukcuoglu, Google’s chief technology officer for AI, said accuracy rises further when Gemini is paired with Google Search. The system can cross-check information online in real time. But experts caution that it still does not guarantee perfect results.
Manos Koukoumidis, co-founder of Oumi, said that users expecting flawless answers may still be disappointed. He believes the technology has progress to make before it can fully replace traditional search behavior.
Rapid Adoption and Student Push
Google is also expanding access. The Gemini app now has 650 million monthly active users, nearly double its March figure. To widen its audience further, Google will again offer its AI Pro subscription free to U.S. college students, saving each subscriber about $240 annually.
Business Applications Gain Momentum
Two dozen companies have already integrated Gemini 3 into their workflows. They are using it to analyze data, translate and transcribe multilingual meetings, organize inboxes, and draft emails. Google expects these capabilities to strengthen its standing in the cloud computing sector, where it competes heavily with Microsoft and Amazon.
Cloud demand is rising. Last month, Google reported a one-third jump in cloud revenue, reaching $15 billion, driven largely by interest in AI-powered tools.
As the global AI race accelerates, Gemini 3 signals Google’s determination to lead the next phase of the industry. The company is betting that stronger performance, wider distribution, and deeper integration into its products will convert curiosity into long-term adoption.
For now, the tech world is watching closely. If Gemini 3 delivers consistent accuracy and meaningful business value, Google could reset the competitive balance. If not, the debate over the true commercial potential of advanced AI will only grow louder. Either way, Gemini 3 has ensured that the race is far from over—and the next moves from Google, OpenAI and Anthropic will shape the future of the entire sector.
