Dvir Ben Ishay
HEBook a call
Book an intro call
Detail 03 · AI news · Google

What's happening at Gemini

Google builds Gemini: the Gemini app, the Gemini API available through Google AI Studio and Vertex AI, and NotebookLM for working with documents and sources. Model announcements mostly appear on the Google DeepMind blog.

Company
Google
Items
21
Updated
Sources

New models

9 items
  1. Model
    Illustration: Google releases Gemini 4 Carbon

    Google releases Gemini 4 Carbon

    Original title: Gemini 4 Carbon is now available

    What it is
    Google has released Gemini 4 Carbon, a new artificial intelligence model. It is designed for complex tasks in cybersecurity, software engineering, and scientific research.
    How it works
    The model uses more data tokens than previous models to reason deeply and run verification loops. It can also coordinate teams of AI agents to work together on large projects.
    Where it is available
    It is available through the API in Google AI Studio, Gemini for business, the Gemini app, Google Antigravity, and AI Mode in Search.
    Example
    A team of these AI agents used planning documents to build an interactive 3D reconstruction of an office building.
  2. Model

    Google Gemini Robotics 2.5 improves robot dexterity

    Original title: Gemini Robotics 2.5 accelerates real-world general dexterity

    What it is
    Google announced Gemini Robotics 2.5, a vision-language-action model. It is designed to improve the dexterity, reliability, and safety of robots.
    How it works
    The model combines vision, language, and action capabilities to help robots perform physical tasks. The source does not provide further details on how it operates.
    Where it is available
    The source mentions Gemini 4 Argon for Robotics October 2026 Models, but does not specify exactly where the model is currently available or applied.
    Example
    Not stated in the source
  3. Model
    Illustration: Google Gemini Omni 1.2 Flash: 15s video for 30% less

    Google Gemini Omni 1.2 Flash: 15s video for 30% less

    Original title: Gemini Omni 1.2 Flash: Fast and cost-effective with longer clips

    What it is
    Google released Gemini Omni 1.2 Flash, a model that creates and edits videos up to 15 seconds long. It is 30% cheaper than the previous version and keeps characters, products, and styles more consistent.
    How it works
    Users can create videos from text, images, or audio, and edit them by typing follow-up instructions in a conversation. They can start with lower-resolution drafts and upscale the final video to higher resolutions.
    Where it is available
    It is available in the Gemini API, Google Cloud, Gemini for business, Google Vids, the Gemini app for Google AI Ultra subscribers, Flow, and Google Ads.
    Example
    A user can combine a photo of a model with front and back shots of three outfits to generate a 15-second virtual try-on video.
  4. Model
    Illustration: Google launches Gemini 4 Carbon, its strongest model

    Google launches Gemini 4 Carbon, its strongest model

    Original title: Gemini 4 Carbon: Our next frontier model

    What it is
    Google announced Gemini 4 Carbon, its strongest AI model designed for cybersecurity and complex, long-term tasks.
    How it works
    The model can run for days to solve multi-step engineering problems, find security vulnerabilities, and write secure code. It is also trained to be more honest and reduce errors or fabricated information.
    Where it is available
    It is currently available to a select group of cybersecurity defenders and trusted partners through the Fairwind Program, with plans to release it to developers and businesses later.
    Example
    The Chrome team used the model to generate high-quality software patches to fix security bugs.
  5. Model
    Illustration: Google unveils Genie 4 for real-time world simulation

    Google unveils Genie 4 for real-time world simulation

    Original title: Genie 4: A new generation of world simulation

    What it is
    Genie 4, a new world model, has been announced. It is designed for real-time simulation.
    How it works
    The model simulates environments in real time. No further details on how it works are provided in the source.
    Where it is available
    Not stated in the source
    Example
    Not stated in the source
  6. Model
    Illustration: Google launches Gemini 4 Argon for robotics

    Google launches Gemini 4 Argon for robotics

    Original title: Gemini 4 Argon for Robotics

    What it is
    Google introduced Gemini 4 Argon, an AI model designed for robotics. It improves physical understanding, code generation, and visual reasoning.
    How it works
    The model plans actions for robots, creates digital simulation twins from videos or sketches, and controls robots in real time. It processes visual data and makes step-by-step decisions to solve physical tasks.
    Where it is available
    The model applies to robotics systems and simulation environments. Specific details about commercial availability, APIs, or regions are not stated in the source.
    Example
    Controlling a four-legged robot to pick up soda cans and drop them into a foot-pedal trash bin after practicing the movement in a digital simulation.
  7. Model
    Illustration: Google releases Gemini 4 Argon, its latest AI model

    Google releases Gemini 4 Argon, its latest AI model

    Original title: Gemini 4 Argon: Built to solve your hardest problems

    What it is
    Google announced the release of Gemini 4 Argon, its latest artificial intelligence model.
    How it works
    The model is designed to solve complex, multi-step tasks in fields like coding, finance, legal work, cybersecurity, and science. Users can interact with it through applications, developer tools, or search.
    Where it is available
    It is available globally through the API in AI Studio and Gemini platform, in Google Antigravity for AI Ultra subscribers, and in the Gemini app and Search AI Mode for Ultra and Pro subscribers.
    Example
    Isomorphic Labs uses the model to analyze complex PDFs and patents to extract information for drug discovery, reducing the process from weeks to hours.
  8. Model
    Illustration: Google launches multimodal EmbeddingGemma 2

    Google launches multimodal EmbeddingGemma 2

    Original title: EmbeddingGemma 2: an open, lightweight multimodal embedding model

    What it is
    Google launched EmbeddingGemma 2, an open and lightweight model for creating mathematical representations of data. The model now supports text, code, images, video, and audio in a single shared space.
    How it works
    It runs directly on personal devices, allowing data processing without sending information to external servers. It is modular, meaning developers can load only the parts they need, such as text only, to save device memory.
    Where it is available
    The model is available for download on Hugging Face and Kaggle, and will soon be on Gemini Enterprise Agent Platform Model Garden. It is designed to run locally on consumer hardware.
    Example
    Finding a specific video clip using a voice memo, or searching through hours of audio recordings using a text query.
  9. Model
    Illustration: Google unveils Gemini 4 Argon for coding and security

    Google unveils Gemini 4 Argon for coding and security

    Original title: Gemini 4 Argon: our next era of frontier intelligence

    What it is
    Google announced Gemini 4 Argon, a new artificial intelligence model designed for software engineering, professional office work, and cybersecurity.
    How it works
    The model can handle complex, multi-step tasks and has a very large limit of 1 million output tokens, allowing it to think deeply and solve difficult problems in one go.
    Where it is available
    It is currently rolling out to a select group of cybersecurity defenders through the Fairwind Program, with plans to expand to developers, businesses, and consumers later.
    Example
    A cybersecurity company named Wiz used the model to find a critical security flaw in healthcare software that exposed sensitive personal information.

Tools and products

4 items
  1. Tool
    Illustration: Google brings Gemini Intelligence to Pixel 10

    Google brings Gemini Intelligence to Pixel 10

    Original title: Gemini Intelligence: New features, on more devices

    What it is
    Google announced new features for Gemini Intelligence and expanded its availability to more devices, including the motorola signature 27 and the Pixel 10 series. The update includes tools for organizing spoken thoughts, automating tasks, creating custom widgets, and getting device support.
    How it works
    Users can turn spoken thoughts into organized text in real-time, ask the assistant to handle tasks like ordering food, or describe a widget in plain words to design it. The system also helps manage phone settings and provides proactive suggestions while driving.
    Where it is available
    The features are available on supported Android devices, including Pixel 10 models and the motorola signature 27. Task automation is expanding to the UK, India, and Japan, and widget creation is coming to the Pixel Watch.
    Example
    If a friend texts you to ask where you are meeting for a pilates class, the system on Android Auto can find your confirmation email and suggest sending the address with a single tap.
  2. Tool
    Illustration: Google and Adobe integrate apps and AI tools

    Google and Adobe integrate apps and AI tools

    Original title: We’re bringing Adobe’s creative and productivity tools closer to your everyday Google apps.

    What it is
    Google and Adobe are expanding their partnership to bring Adobe's creative and editing tools into Google apps, and Google's AI models into Adobe's apps.
    How it works
    Users will be able to access Adobe Acrobat directly from Google Drive and Gmail. Businesses will also be able to use Adobe's AI assistant inside Gemini Enterprise to work on marketing and design tasks.
    Where it is available
    The integrations will be available in Google Drive, Gmail, Gemini, Gemini Enterprise, Google Search, and Photoshop.
    Example
    A business team can use the CX Enterprise Coworker agent inside Gemini Enterprise to design graphics and launch a marketing campaign.
  3. Tool
    Illustration: Google Antigravity speeds up drug discovery

    Google Antigravity speeds up drug discovery

    Original title: Agentic cheminformatics with Google Antigravity

    What it is
    Google introduced Google Antigravity and Cheminformatics Skills to speed up drug discovery processes.
    How it works
    These tools work by searching for chemical compounds and proposing new structures called rigidified macrocycles.
    Where it is available
    Not stated in the source.
    Example
    The system can be used to speed up the search for antimicrobial peptides.
  4. Tool
    Illustration: Google Antigravity speeds antimicrobial peptide search

    Google Antigravity speeds antimicrobial peptide search

    Original title: Accelerating the search for antimicrobial peptides with Google Antigravity

    What it is
    Google demonstrated how its development platform, Google Antigravity, helps scientists search for new antimicrobial peptides to fight antibiotic resistance. The company also introduced a built-in 3D visualization tool called Biomolecular Structure Viewer.
    How it works
    A researcher and an AI agent work together in a single workspace. The agent handles tasks like searching databases, analyzing protein structures, and summarizing scientific literature, while the researcher guides the overall strategy.
    Where it is available
    The tools are available within the Google Antigravity development platform.
    Example
    Researchers used the platform to find a new candidate peptide to fight bacteria that cause respiratory tract infections, and this candidate is now being tested in a physical lab.

Behind the scenes: company and research

7 items
  1. Company
    Illustration: Google Cloud launches Gemini for healthcare

    Google Cloud launches Gemini for healthcare

    Original title: Google Cloud launches Gemini for healthcare.

    What it is
    Google Cloud launched Gemini for healthcare, an AI solution designed to connect clinical, operational, and administrative systems. It helps healthcare teams automate workflows and reduce administrative paperwork.
    How it works
    The platform connects with existing third-party systems like electronic health records, billing platforms, clinical guidelines, and Medicare datasets. It uses this information to run complex, multi-step tasks with human oversight.
    Where it is available
    It is initially available in preview with launch partners including HCA Healthcare, Highmark Health, and Northwell Health.
    Example
    The system can be used for medical records synthesis and prior authorization review.
  2. Company
    Illustration: AlphaProof Code writes 3M lines of proofs for Signal

    AlphaProof Code writes 3M lines of proofs for Signal

    Original title: Introducing AlphaProof Code: a breakthrough in verifying complex software at depth

    What it is
    Google DeepMind introduced AlphaProof Code, which are AI agents for software verification, and CodeMender, an AI agent for code security.
    How it works
    The agents wrote three million lines of proofs in the Lean programming language to verify that the software works correctly.
    Where it is available
    This was applied to Signal's core encryption. Other availability details are not stated in the source.
    Example
    Verifying the core encryption of the Signal messaging application.
  3. Research
    Illustration: Google unveils Gemini 4 Argon and Carbon

    Google unveils Gemini 4 Argon and Carbon

    Original title: The latest AI news we announced in October 2026

    What it is
    Google announced new AI models, including Gemini 4 Argon for complex tasks and Gemini 4 Carbon for cybersecurity and engineering. They also introduced Genie 4 for creating interactive simulations and updated security features for Android.
    How it works
    Gemini 4 Argon is available for tasks like coding and financial analysis. Gemini 4 Carbon is currently in pre-release testing with the U.S. government and selected defenders to refine its safety features.
    Where it is available
    Gemini 4 Argon is available globally across Google AI Studio, the Gemini API, Google Cloud, Search, and the Gemini app. Gemini 4 Carbon is in limited pre-release testing.
    Example
    A user can use Gemini 4 Argon for complex coding, financial analysis, or legal workflows.
  4. Company
    Illustration: Google and Singapore build Gemma scam detector

    Google and Singapore build Gemma scam detector

    Original title: Google DeepMind and Singapore Government collaborate to build fine-tuned AI model for scam detection

    What it is
    Google DeepMind and the Singapore Government are collaborating to build a fine-tuned AI model to detect scams.
    How it works
    The system uses lightweight AI models on the user's device to identify scam attempts in private chats without sending any data to the cloud.
    Where it is available
    The technology runs directly on the user's device for private chats.
    Example
    Intercepting conversational scams within private chat applications.
  5. Research
    Illustration: Google begins broad rollout of its Gemini 4 Argon AI model

    Google begins broad rollout of its Gemini 4 Argon AI model

    Original title: Developing Gemini 4 Argon responsibly and safely

    What it is
    Google DeepMind is starting to broadly roll out its Gemini 4 Argon AI model and has shared details about its safety and security approach. The model is designed to handle complex, real-world tasks in fields like coding, finance, cybersecurity, and science.
    How it works
    It uses three layers of real-time protection to screen inputs, analyze internal signals for harmful intent, and continuously monitor responses. It is also trained to respect boundaries, ask for human confirmation, and use an automated monitor to prevent it from taking unauthorized shortcuts.
    Where it is available
    The model is beginning to roll out broadly, and Google is expanding its Fairwind Program to more trusted cyber defenders.
    Example
    The security company Wiz used Gemini 4 Argon to find a critical bug in hospital software, helping protect hundreds of thousands of patient records.
  6. Research
    Illustration: Google launches Gemini 4 Argon and Windows app

    Google launches Gemini 4 Argon and Windows app

    Original title: The latest AI news we announced in September 2026

    What it is
    Google announced several AI updates, including a new model called Gemini 4 Argon and new voice features. They also added integrations with popular apps and released a dedicated Gemini app for Windows.
    How it works
    Gemini 4 Argon uses advanced reasoning to solve complex problems. Users can connect apps like Airtable and monday.com directly inside Gemini to complete tasks without switching tabs.
    Where it is available
    Gemini 4 Argon is currently rolling out to early testers in the Fairwind Program. Other features, like the Windows app and new voice tools, are available in Google Workspace and Google AI Studio.
    Example
    A user can use their voice to manage tasks in Gmail and Docs, or use Gemini on Windows to brainstorm ideas using a keyboard shortcut.
  7. Company
    Illustration: Google unveils SynthID Bio to watermark AI proteins

    Google unveils SynthID Bio to watermark AI proteins

    Original title: Introducing SynthID Bio

    What it is
    SynthID Bio was introduced as a proof of concept for watermarking proteins generated by artificial intelligence. This technology aims to watermark these proteins while keeping their biological function intact.
    How it works
    The technology applies a watermark to AI-generated proteins. It does this in a way that preserves the biological function of the proteins.
    Where it is available
    Not stated in the source.
    Example
    Not stated in the source.

Money, IPOs and deals

1 item
  1. BusinessTechCrunch
    Illustration: Google expands AI Search ads globally

    Google expands AI Search ads globally

    Original title: Google expands AI Search ads globally as core ad growth slows

    What it is
    Google is expanding advertisements within its AI-powered Search features to international markets. The company is also introducing multi-turn ads in AI mode within the United States.
    How it works
    Advertisements are placed inside AI-generated overviews and answers. These include sponsored product listings, recommendations, and conversational ads that respond to questions.
    Where it is available
    The expansion applies to international markets globally, with specific features like multi-turn ads launching in the United States.
    Example
    A user might see sponsored product listings or recommendations integrated directly into the answers generated by the AI.
Every morning · in Hebrew

Yesterday's AI, in this morning's inbox

The new models, tools and deals, each with a short note on what it changes for a business.

How often?

A short issue in the morning, on every day there is news. Written in Hebrew.

Unsubscribe in one click. Your email is used only for the brief.

Past issues