Google's October AI recap: Gemini 4 Argon, Carbon and Genie 4

Gemini 4 Argon rolls out, Carbon enters pre-release testing

Google describes Gemini 4 Argon as a frontier reasoning model. After early safety testing, it rolled out globally for complex coding, financial analysis and legal workflows. It is available across Google AI Studio, the Gemini API, Google Cloud, Search and the Gemini app. Google also says it powers physical-world reasoning through Gemini 4 Argon for Robotics.

Two weeks later, the company announced Gemini 4 Carbon, which it calls its strongest model yet, with state-of-the-art cybersecurity and long-horizon capabilities. Carbon is not yet widely available. It is in pre-release testing with the U.S. government and a subset of trusted cyber defenders via the Fairwind Program. Google says it is refining safety guardrails and expanding partner testing ahead of a broader launch to developers, businesses and consumers.

Genie 4 and tools for building

Genie 4 is Google's new world model for real-time interactive simulation. According to Google, it brings realism, advanced physics, lifelike human movement and native audio, and users can steer scenes as they unfold using natural language. The company says it advances its ability to create safe training environments for robots and AI agents, which it calls a crucial step toward AGI. Demos are on the Google DeepMind blog.

Google also introduced Playground, an experimental gaming platform where anyone can build, remix and share single- and multiplayer games from text prompts without coding. Android users can now build personal AI apps with Google AI Studio and install them through Google Play, and music producers can create custom instruments and audio plugins in Google Flow Music.

For developers, Google released EmbeddingGemma 2, an open embedding model built on the Gemma 4 architecture with 740 million parameters. It maps code, images, video and audio into a single space on everyday devices. Its modular design cuts local vector storage by up to 6x, according to Google, which should help developers build private, offline apps.

Safety, verification and health

Google is expanding SynthID Detector globally in English so anyone can check whether an image, video or audio clip was generated with AI. The company says it has watermarked more than 180 billion images and videos and 240,000 years of audio with SynthID. Built-in verification tools in Search, Chrome and the Gemini app handle over a million requests each day, and the detector also identifies AI content from partners including OpenAI, NVIDIA, Kakao and Apple.

On Android, AI-powered Scam Detection in Phone by Google and Google Messages already flags more than 5 million suspected scam calls and texts every month. Google is now adding an opt-in Enhanced protection mode in Google Messages. It uses Private AI Compute and the latest cloud Gemini models to catch sophisticated conversational scams, with expanded chat alerts and stronger call-forwarding and screen-sharing defenses.

In health, Google launched the Fitbit Edge fitness tracker, available now, with up to seven days of battery life, precision GPS and automatic workout recognition across 50+ activities. It pairs with the Gemini-powered Google Health Coach. Google also rolled out more than 55 updates across Google Health and Fitbit Air.

In its first publication in the main journal of The Lancet, researchers from Google and Beth Israel Deaconess Medical Center evaluated the AMIE conversational medical AI system in real-world primary care visits. By gathering patient histories and generating clinical summaries before urgent care appointments, AMIE helped clinicians prepare in 75% of cases and matched doctors' final differential diagnoses 90% of the time.

Education, space and research

For World Teachers' Day, Google highlighted how teachers use Gemini to personalize lessons, and launched a global idea search with TED. Students can use Gemini Notebook for interactive learning overviews and custom quizzes, plus new co-creation controls to shape Slide Decks and Video Overviews before they generate.

Google's first Project Suncatcher prototype satellite returned the first AI response generated by a data center-class Google TPU in space: a haiku composed onboard by Gemma 4. Google says this confirms the hardware survived launch and can run AI workloads in orbit. The team will keep testing how the chips handle radiation and thermal extremes. The recap also mentions new research shared at the Research@ 2026 event.

What it means

The roundup shows Google shipping models and consumer features at a fast pace while staging access to its most capable system. Argon is broadly available now, while Carbon is limited to government and vetted defenders for the time being. For everyday users, the most immediate changes are the opt-in scam protection in Google Messages, the wider SynthID Detector and the new Fitbit Edge. Developers get an open multimodal embedding model and no-code routes to building apps and games.