Google just leapfrogged every competitor with mind-blowing AI that can think deeper, shop smarter, and create videos with dialogue

Google Announced a sweeping set of artificial intelligence development annually on Tuesday I / o developer conferenceTo present stronger AI models, expand the search capabilities and launch new creative means that push the boundaries of its technology.

The company based on a mountaine appeared Gemini 2.5 Accessoriesrolled AI mode in search of Presented to all US users New generative media modelsand a prize called 249.99 Monthly subscription steps Google has ultra For power users – all reflect all that accelerates the AI ​​speed along the product ecosystem.

“More intelligence, everywhere, everywhere, the world, which means what progress is in a new stage of the AI ​​platform change in decades for decades for decades, enterprises and communities for decades.”

Developed Thought: Gemini 2.5 Models provide revolutionary “deep think”

The center of Google has a continuous evolution Gemini Great Language ModelsWith significant developments for both Projector and Flash versions. Updated Twins 2.5 Flash In general, after a while, it will be available in general with Pro.

Most noteworthy “Deep thinking“An advanced mode of thinking for the Pro model that Google claims, gives the performance of progress in complex positions using parallel thinking techniques.

“Deep thinking adapts the basic performance to its limits and conveys groundbreaking results,” he said. “This is an effective result in Usamo 2025, one of the most difficult mathematical criteria. Also, a criterion for competitive level coding, leads to Livekodebeng.”

The company continues with caution Deep thinkingIt plans to submit to reliable testers for feedback before leaving first. This measured approach focuses on the accommodation of Google’s responsible for the border opportunities that push the borders of the AI.

REIMAGINGING Search: AI mode expands with personalization and agent features

Google brings AI to a deeper basic search product, rolls “AI mode“After previously restricting it to laboratory testers, all US users. This uses a technique called an alternative search experience”query fan“To break down questions to subtopes and release numerous searches than traditional search.

“The AI ​​mode is a more advanced thinking and more powerful AI search with multimomhood and multimodality and the ability to deepen the Internet via useful links,” said Liz Reid, VP and Google search head.

The company discovered the existing effective dimensions AI Views Now the feature that reaches more than 1.5 billion users. “In our largest markets like the United States and India, AI Views are more than 10% of Google’s use for queries that show AI views,” Reid said.

Includes new features that come to AI mode Deep search Comprehensive research reports, live opportunities for real-time visual assistance and personalization options that users can enter information from Google accounts. This personalization requiring open user choice aims to achieve more relevant results by understanding individual advantages and contexts.

Google gives a significant impetus to the EU employed shopping experience Virtual attempt feature This allows users to see how the clothes will look only using a single photo. Technology represents the main development of easier and personalization in online shopping.

“This is a situation I have found perhaps my likes and I find five clothes and I see how it looks in the models there.

The system is equipped with a special image generation model designed specifically for fashion applications. According to Srinivasan, “3D-shaped” and pieces “a very deep concept” and the fabrics allow you to realize how different types of body species fit.

Out of the visual test, Google is also applied Agentic Checking When items reach the price point specified by the user, automatically can complete the purchases. This feature manages the entire audit process Google PaymentTo make Google’s daily duties facilitate, the agent demonstrates how it applies the AI ​​capabilities.

Google has introduced significant updates to generative media models 3 I see Video generation and Fig. 4 For pictures. The most dramatic progress is to create videos with Velo 3 synchronized audio, including environmental sounds, effects and character dialog.

“For the first time, we arise from the silent period of the video generation,” Hassabis said. “Only VEO 3 does not offer more surprising visual quality, but can also create sound effects, background sounds and even dialogs.”

This advanced models strength FlowNew to Google’s AI Film Boots Designed for creative professionals. Stream combines the best AI models of Google to help create movie clips and scenes with more intuitive interface of storytellers.

“Streaming, when time is slow down and can be diligent, iteratia and probably,” he inspires himself. The tool has already been tested with several directors who created short films using technology in traditional ways.

Fig. 4At the same time, the quality of image quality, improvement of the quality of the image by providing special attention to the indication of typography and text – makes it particularly valuable to create marketing materials, presentations and other content.

Immersive Communications: Google Ray is developing from the project starline

The company has announced that Project star lineA few years ago, experimental 3D video communication technology, demonstrated for the first time, becomes a commemorative trade product Google beam. This technology creates a sense of being in the same room as someone communicates remotely.

“Google Beam will be a new AI-first video communication platform,” said Pichai said. “Beam uses a new artistic AI video model that turns video streams into real 3D experience.”

The system uses a number of cameras to capture different aspects of participants, then combine these streams and display them on the 3D light area screen with accurate head tracking. The result is a deep immersive conversation experience that goes beyond the traditional video call.

Google partnered with HP to bring the first Google beam At the end of this year, the devices to be marketed for selected customers. Technology also presents the ability of the language and expression of language, which allows the quality and expression of language barriers and phrases, which protects the sound quality and phrase – Google will also meet.

Premium Introduction: New ultra subscription levels target Tier power users and specialists

To make money the most advanced AI offers, a premium subscriber called Google Google has ultraPrice for $ 249.99 per month. This level provides early access to Google’s most skilled models, the highest use limits and experimental features.

“You require a director, developer, creative professional or simply the best of Google AI

Includes access to Ultra Plan Veo 3 with sound generation, Deep thinking mode When available, File shooting, Project Mariner Agency capabilities and 30TB storage. Also accumulates with YouTube Award.

“The way to think about Google AI Ultra is almost your VIP access to all of the Google’s AI. Therefore, we put the highest level limitations and the products and features there,” said Josh Woodward, Google Labs and Gemini VP.

Google’s Standard AI Pro subscription will last $ 19.99, resulting in this more affordable choice.

The study faced with reality: Google has AI vision form

Google I / O ads reflect a company that successfully changed the products of how people interact with technology, successful research investments. The AI ​​- ancestry signal an evolutionary evolutionary evolution outside the current generation of EU – an AI – AI – an AI – an AI – an AI-aid EU.

“One of the things I found in magic is especially wanted … People simply adapt to the power of possible,” Pichai said. “I think the great thing that people are excited, when you do [interaction] is more natural and intuitive. “

Google’s expansion ecosystem for enterprises and developments that attract AI strategies offer strong tools, but integrated ways, expenses and data confidentiality requires careful consideration. The inclusion of key products in the main products of the company offers a long-term strategy to protect the existing markets and creates new income streams.

As these technologies move from laboratories to the daily use, Pichai’s current AI momentary observer stresses: We are witnessing the transformation of polite opportunities to practical means that people work and communicate naturally. Race to build aids AI not only relevant to the technical capacity – it consists of bringing less to the moments we need to feel less like using technology and thus understanding it.