Google、7月のAI更新を公開：スケーラブルなエージェント用Geminiモデル3種とロボット向けER2を発表

Aug 04, 2026

|

9 min read

Here’s a recap of some of our biggest AI updates from July, including three new Gemini models for building AI agents at scale and Gemini Robotics ER 2 to help robots reason, collaborate, and solve real-world tasks.

News from Google Team

Your browser does not support the audio element.

Listen to article

[[duration]] minutes

This content is generated by Google AI. Generative AI is experimental

For more than 20 years, we’ve invested in machine learning and AI research, tools, and infrastructure to build products that make everyday life better for more people. Teams across Google are working on ways to unlock AI’s benefits in fields as wide-ranging as healthcare, crisis response, and education. To keep you posted on our progress, we're doing a regular roundup of Google's most recent AI news.

Here’s a look back at some of our AI announcements from July.

July brought a wave of AI advances to support developers, simplify everyday life, and protect local communities. We gave developers faster, more efficient Gemini models for scaling production agents and advancing robotics. We also expanded our creative suite with next-generation music and video tools, and made Search and Gemini Spark even more connected to your go-to daily apps. Plus, we announced a new skilled trades alliance and new satellites for early wildfire detection — and Google DeepMind founder Demis Hassabis shared his vision for this critical moment in history. Together, our July updates showed how AI can be more practical, creative, and helpful than ever before.

Check out three new Gemini models. Developers and customers building production AI agents need higher token efficiency, lower latency, and more reliable performance. Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber are built to meet the sweet spot of efficiency and quality to enable scaling agentic workflows.

See how robots gear up with Gemini Robotics ER 2. We launched Gemini Robotics ER 2, our most capable "embodied reasoning" model yet. Built for developers crafting genuinely helpful machines, ER 2 bridges the gap between digital smarts and the physical world. It lets systems chat naturally with people, make sense of their surroundings, and work through complex, multi-step tasks.

Get started with Gemini Intelligence on the latest devices from Galaxy Unpacked 2026. With the launch of the Samsung Galaxy Z Fold8 Ultra, Fold8, and Flip8, we’re making it easier than ever to get started on Android while also rolling out the first capabilities of Gemini Intelligence.

1

Switch to Android even more easily and bring your data with you. With the launch of Samsung’s new Z Series at Galaxy Unpacked, we introduced a new native migration experience built directly into Android 17 that lets you transfer even more data types from an iPhone wirelessly — without downloading a separate app.

Streamline your research with Gemini Notebook. Gemini Notebook is the same standalone product as the hugely popular NotebookLM, only now it's doing more across the Google ecosystem — including inside the Gemini app and Google Search — and it’s updated with a secure cloud computer.

Explore the first iteration of the AI & Economy ATLAS. Google’s ATLAS is an expansive look at how people are using AI at work and in day-to-day life. This ongoing, large-scale, de-identified study is designed to offer an early view of a quickly moving landscape: AI’s capabilities are advancing, its use is evolving, and tools for observing its impact on the economy are still a work-in-progress.

Meet the Alliance for America’s Skilled Trades. To help meet the growing demand for skilled trades workers, we’re joining forces with BlackRock, Carhartt, and Ford to launch the Alliance for America’s Skilled Trades. Together, we’ll build a stronger pipeline for the skilled trades, scale evidence-based training approaches, and expand opportunities nationwide through cross-industry partnerships.

See how AlphaEvolve can help Cloud customers solve their hardest problems. Finding the most efficient algorithm can be challenging. To help, we made AlphaEvolve, our Gemini-powered AI code-optimization agent, generally available to all Google Cloud customers on Gemini Enterprise Agent Platform. AlphaEvolve acts as an evolutionary collaborator: You provide a baseline algorithm and your goals, and it automatically searches for better solutions, returning human-readable, optimized code.

Connect more of your apps to Search. You can already connect some of your favorite services to the Gemini app. To help you get even more done, we’re enabling you to securely link and interact with your go-to apps like YouTube Music directly in AI Mode.

With Personal Intelligence and your connected apps, Search can give even more tailored responses.

Handle complex web errands with Gemini Spark. We expanded Gemini Spark to more users around the world, and added a new feature to help you handle complex web errands. Now, with your permission, Spark can use your logged-in accounts and saved passwords to handle tedious web errands, like scheduling viewings for apartments or researching flight options and starting the booking process.

Create and personalize your videos in Google Vids with new updates. Two new features — Gemini Omni and personal avatars — make video creation easier than ever. Now you can generate and edit high-quality clips using everyday language, and create a custom digital avatar to star in your videos.

See how we reconstructed Pelé’s “lost” goal. On August 2, 1959, Pelé scored the most beautiful goal of his career: three consecutive “sombreros” over defenders and the goalkeeper without the ball ever touching the ground. But the moment was never captured on film. We used Google DeepMind technology to reconstruct a lost piece of football history and tell the story of how it all came to life.

Celebrate 25 years of visual search innovation with new features. It’s been 25 years since we launched Google Images, beginning the evolution of Search from a text-based experience to one that allows you to visually explore the world. To celebrate, we announced a new browseable home for Google Images and image generation through AI Overviews in Search. We also looked back at the major launches that transformed how we search with — and for — images.

Make music with the newest music generation model, Lyria 3.5. Google Flow Music helps you create songs you love with total creative control. Powered by our latest Lyria 3.5 model, it delivers significant advancements in musicality, lyrics, and vocal quality to help you craft richer tracks.

See how NOAA and Google Cloud are advancing weather forecasting. NOAA (National Oceanic and Atmospheric Administration) is modernizing its operational supercomputing system by using Google Cloud’s high-performance H4D virtual machines to run its atmospheric models. As one of the first organizations to bring numerical weather prediction directly to the public cloud, NOAA provides meteorologists with the scale, speed, and AI capabilities needed to run massive simulations and deliver faster, life-saving early warnings.

Learn how AI breakthroughs can help people prepare for and respond to natural hazards. With the release of the UN report “Leveraging AI to enhance multi-hazard early warning systems,” we shared how our teams at Google support crisis resilience efforts with advanced AI-based breakthroughs, detection, and forecasting — while working with partners to make timely information available to billions of people. Our goal is a world where no one is surprised by a natural disaster.

Discover how three new satellites joined the fight against wildfires. Three new FireSat satellites successfully launched from Vandenberg Space Force Base in California, expanding a global initiative led by the Earth Fire Alliance (EFA) with Google Research to help detect wildfires before they spread.

Examine the bioresilience program designed to protect global communities. Google DeepMind and Isomorphic Labs shared a joint approach to bioresilience. The goal is twofold: to prevent threat actors from misusing our models, and to ensure that governments, scientists, biosecurity experts and our teams can harness these technologies to build a more resilient world.