By Kenrick Cai
SAN FRANCISCO (Reuters) -Alphabet’s Google on Tuesday launched the latest version of its artificial intelligence model Gemini, emphasizing that the new capabilities will be immediately available in several profit-generating products like its search engine.
Gemini 3, arriving 11 months after the second generation of the model, appears on paper to keep Google at the forefront of the AI race. During a press briefing, executives highlighted Gemini 3’s lead position on several popular industry leaderboards that measure AI model performance.
CEO Sundar Pichai described it as “our most intelligent model,” in a company blog post.
However, the AI race has increasingly shifted away from benchmarks to money-making applications of the technology, as Wall Street watches for signs of an AI bubble. Alphabet’s stock has so far been buoyed this year largely due to the financial success from AI offerings from its cloud computing division.
But even with leading developers like Google, OpenAI and Anthropic behind them, new AI model updates have had trouble distinguishing themselves, only attracting attention when they fail, as Meta experienced earlier this year.
Google emphasized that Gemini 3, unlike past releases, was already underpinning a handful of revenue-generating consumer and enterprise products at launch.
“We think Gemini has set quite a new pace in terms of both releasing the models, but also getting it to people faster than ever before,” Koray Kavukcuoglu, Google’s chief AI architect, told reporters during the briefing.
Pichai said the Gemini 3 launch marked the first time that Google had incorporated its new model into its search engine from day one. In the past, new versions of Gemini took weeks or months to embed into Google’s most highly used products.
Paying users of Google’s premium AI subscription plan will have access to Gemini 3 capabilities in AI Mode, a search feature that dispenses with the web’s standard fare in favor of computer-generated answers for complicated queries.
NEW FEATURES
Improvements to Gemini 3 in domains such as coding and reasoning enabled Google to build out a set of new features, both for consumers and enterprise customers.
The company debuted “Gemini Agent,” a feature that can complete multi-step tasks, such as organizing a user’s inbox or booking travel arrangements. The tool brings Google closer to its AI chief Demis Hassabis’ vision for a “universal assistant” that has been referred to internally as AlphaAssist, as Reuters previously reported.
Google also redesigned the Gemini app to return answers reminiscent of a full-fledged website, a further blow to content publishers who rely on web traffic to generate revenue.
Josh Woodward, the vice president in charge of the app, demonstrated to reporters how Gemini can now respond to a query like “create a Van Gogh gallery with life context for each piece” by generating an on-demand interface with visual and interactive elements.
For business customers, Google previewed a new product called Antigravity, a new software development platform where AI agents can plan and execute coding tasks on their own.
(Reporting by Kenrick Cai; Editing by Kenneth Li and Himani Sarkar)
Comments