Studio-grade 3D fur dynamics, running live in your browser with Gemini 4 Argon. In @Antigravity, Argon coded a custom WebGL2 fur engine with automatic lightweight fallback — using GPU groom shaders, procedural character meshes, and animations for interactive, real-time hair physics.
36
13
130
39,409
With Gemini 4 Argon, any builder can transform physical spaces into interactive games 🏁 We snapped a photo of a real desk and used Argon to orchestrate subagents, Blender MCP, and Nano Banana to transform everyday objects into this playable 3D track.
160
153
1,517
707,809
Gemini 4 Argon can turn messy data into clean, visual narratives in just minutes 🪐 Here’s how it processed 80,000 files from @NASA's open archives (CSVs, images, PDFs, web articles, XML logs, and raw videos) to build a Google Slides deck based on strict design guidelines.
135
153
1,524
187,232
At @IsomorphicLabs, researchers are using Gemini 4 Argon as a partner to accelerate drug discovery and design research 🧬 Argon quickly synthesizes complex PDFs and patents to extract pertinent information, reducing a process that typically takes weeks into just hours.
62
71
630
176,905
With Gemini 4 Argon, your research has new ways to tell its story 📖 We used Nano Banana, Veo, and Argon to turn a dense academic PDF into a full-color, animated storybook in the @GeminiApp.
151
327
3,984
679,527
Our researchers also used Argon in @Antigravity to optimize xTB, one of the most-cited open-source quantum chemistry packages for predicting molecules' 3D shapes and screening drug and materials candidates. They made xTB up to 2.3x faster at predicting molecular structures and 1.8x faster at conformer search, contributing the changes to the official xTB repository on GitHub → goo.gle/4s6dko0
1
1
30
7,162
For Google's open-source Agent Development Kit (ADK), the team gave Argon a skills rulebook and one goal: eliminate framework overhead without breaking existing apps. Working autonomously overnight, Argon surfaced optimizations that cut in-memory session prep time by 40% in long sessions and sped up cold imports by 21%. See it live in google/adk-python → goo.gle/4eAtLDA
3
37
7,439
Software engineers at Google using Argon in @Antigravity are shipping high-quality code faster than ever. Over the past few weeks, agentic code submissions increased by more than 35% — while cutting code rollbacks by nearly a third. Read more about the impact of Gemini 4 Argon in Google Antigravity → goo.gle/4hPJbGo
2
3
63
18,927
In @GoogleChrome, Gemini 4 Argon helped our team separate core webpage handling from a 6-million-line legacy folder of browser features in Chromium’s 40-million-line codebase — completing months of work in just weeks. The result for Chrome's billions of users: eliminating an entire category of browser crashes, and letting live webpages move fluidly between tabs, side panels, and floating windows without reloading.
11
9
113
13,663
Teams across Google put Gemini 4 Argon to work on our hardest engineering problems. Here’s what happened 🧵
133
123
1,890
995,809
View the source code → goo.gle/4jYmj8R
11
4
62
30,871
We asked Gemini 4 Argon to make a realistic 3D scene in pure Python with zero external libraries. It wrote 5,000+ lines of code and built its own 3D ray-tracing engine, custom video tool, and audio mixer to render and document this sunlit library from scratch — producing, narrating, and assembling this video with Gemini 3.8 TTS and music by Lyria 3.5.
118
126
1,362
177,420
From optimizing codebases to creative writing, here's how @TulseeDoshi and other Google teams are using Gemini 4 Argon 👇
33
18
248
116,542
Today, at our Research@ event, we shared research breakthroughs that can help amplify human ingenuity. From agents that can help solve real-world problems, like public health threats and environmental crises like extreme flooding, to models that can aid advanced medical workflows and improve doctor-patient relations, our latest updates will help advance science and technology. Here are some recent highlights.
34
30
240
64,069
We used Gemini 4 Argon in @GoogleAIStudio to build an interactive 3D island sandbox 🏝️ Every brush stroke simulates living terrain, sand, and water physics in real time using custom Three.js and GLSL shaders.
83
114
1,621
251,275
We taught a robot to play cornhole using Gemini 4 Argon 🦾 Just by watching a video of humans playing, Argon mapped the physics, coded its own simulation to practice, and taught @Apptronik's Apollo humanoid to sink 7 shots in a row.
85
112
1,099
156,767
Gemini 4 Argon began rolling out today. See how devs are already putting Argon to work ↓
139
275
3,359
306,759
Our engineers used Gemini 4 Argon to eliminate an entire category of browser crashes in @GoogleChrome. Argon helped separate core webpage handling from a 6-million line legacy folder in Chromium's 40-million line codebase, completing months of work in just weeks. Here's how ↓
151
260
3,524
925,210
Gemini 4 Argon is available now for Google AI Ultra and Pro subscribers in @GeminiApp and AI Mode in Search, Ultra subscribers in @Antigravity, and to developers and enterprises via the Gemini API in @GoogleAIStudio and the Gemini platform. Argon will launch at an introductory price of $2 per million input tokens and $10 per million output tokens, with cached input tokens priced at 95% off input token price.
43
87
583
136,065
In robotics, Gemini 4 Argon is speeding up real-world physical learning. We let Argon watch a video of a human playing cornhole. It mapped the environment, calculated whole-body joint movements, and autonomously built the simulation code to practice. When transferred to Apptronik’s Apollo humanoid, the robot sank seven straight shots in a real-world trial.
29
138
1,116
224,191
Argon brings major improvements in spatial reasoning, mathematical precision, and visual coding. Using multi-agent loops in @Antigravity, it autonomously built this entire pagoda complex, sculpting the architecture, foliage, and lighting from scratch in Blender.
19
24
480
55,572
Today, Gemini 4 Argon is officially rolling out.
474
1,147
10,858
1,180,866
183
367
5,251
988,216
Introducing Fitbit Edge, our next-generation health and fitness tracker. Featuring sleek design, a brilliant display, glanceable fitness guidance and advanced health tracking to support your wellness goals. It's designed for both active lifestyles and everyday wear.
141
220
2,363
1,006,704
This week, we’re celebrating World Space Week with a series of Google Doodles spotlighting stunning views of the universe, honoring the technology and human ingenuity that make space exploration possible. 🚀🌌Happy Space Week! #GoogleDoodle
80
99
997
193,318
Here’s how to get the lowest possible price on your holiday flights, according to five years of aggregated Google Flights data. 👇 ✈️ Connecting flights can also be a big money saver. Nonstop tickets are 20% more expensive on average. ✈️ When you travel is key for saving: Departing on a Monday and returning on a Tuesday or Wednesday saves about 14% over weekend flights (and up to 20% domestically versus Sunday). ✈️ Wednesday beat Tuesday as the absolute cheapest day to book a flight, but the difference is negligible — buying a ticket on a Wednesday only saves you about 1.4% compared to Sunday, the most expensive day.
61
88
624
128,746
Our medical research system, AMIE, is the first patient-facing conversational diagnostic tool of its kind to be studied prospectively in a real-world clinical setting. In a study, published today in @TheLancet, we found that when patients chatted with AMIE before an in-person appointment, the interaction built their confidence and helped them organize their thoughts. And physicians who reviewed the information from the patient conversation got more time back to focus on collaborative care and shared decision-making instead of digging through data.
57
102
609
145,465
In our prospective clinical study, our models were able to pinpoint the gestational age within 4 days of accuracy which is a tight enough zone that can have a really meaningful clinical impact. If we're able to expand these tools to low resource settings, we can start to move the needle on maternal deaths and bridge that gap in care that we see all over the world.
4
9
65
29,159
Around two thirds of people around the world lack access to diagnostic health tools, like ultrasounds. We’re working to change that. Here’s how ⬇️
52
70
364
866,078
What are people craving around the world? Our first-ever @GoogleMaps Fan-Favorite Dining List taps into Maps reviews, ratings, and searches for directions to analyze the biggest dining trends and popular restaurants across 10 global food hot spots: Atlanta, Chicago, London, Melbourne, Montréal, New York, San Francisco, Santiago, Seattle, and Tokyo.
47
44
361
865,191
Some things can only be tested in space — like how the hardware holds up to solar events, cosmic rays, or bitflip errors caused by radiation — which is why we’re putting our first TPUs in orbit.
11
13
115
32,269
Project Suncatcher is our moonshot exploring whether we can one day host machine learning infrastructure in space 🚀 To do so, we need to know whether our AI hardware can operate in orbit. Can our TPUs handle the physical stress of spaceflight and the radiation and thermal extremes of space? These are the questions Google researchers have been exploring for the past few years. Last week, we launched our first test satellite carrying four TPUs into orbit — putting us one step closer to getting answers.
152
260
1,936
1,165,549
Want to check if an image, video, or audio file was made using AI? Here’s how: 1️⃣ Go to synthid.com and upload an image, video, or audio file 2️⃣ The portal scans the media to detect if the file contains a SynthID watermark from Google or our partners
15
58
332
57,723
Today, we're expanding SynthID Detector in partnership with @OpenAI, @NVIDIA, Kakao & soon @Apple as part of an industry-wide effort to make AI-generated content more transparent. Available globally in English, our verification portal lets you check if media was made with AI 🧵
265
882
5,788
1,067,348
During the ongoing Ebola outbreak in the Democratic Republic of Congo (DRC), our partners at @WHOAFRO put Google Earth AI to work to uncover transmission blind spots faster. They used our Geospatial Reasoning agent prototype to map areas where exposure risk and human mobility are high. In minutes, the team pinpointed 48 exposed settlements and located more than 45,500 at-risk people — a process that normally would have taken weeks.
3
4
53
27,395
Google Earth AI can bridge critical gaps across diverse geographies, diseases, and areas of public health using autonomous predictions and population dynamics. The system combines environmental signals and other data sources with AlphaEarth Foundations and our Population Dynamics Foundation Model (PDFM), and a prototype Geospatial Reasoning agent. In plain language, researchers ask questions like where a disease is likely to spread next, and the AI automatically gathers relevant models and datasets to build an accurate prediction model. By combining satellite views with population patterns, the tool reveals hidden risk factors, helping identify issues earlier.
10
9
63
14,001
Forecasting where a disease outbreak might spread usually requires weeks of stitching together fragmented satellite imagery, census tables, and local health records. Now, new research shows how Google Earth AI can help communities address public health crises more quickly and proactively. By combining insights beyond simple statistics — like behavioral trends, geospatial AI models, and more — public health teams can better understand complex public health issues, bridge reporting gaps, and shift emergency response from reactive management to proactive prevention.
37
34
301
128,204
Best-in-class for its size at 740M parameters, EmbeddingGemma 2 outperforms some models over twice its size while using from as little as ~191MB to 567MB of active RAM. It also features a 4x larger 8K context window than the first generation, processing up to 5.5 minutes of audio, 29 images, or 58 video frames in one pass.
6
3
69
26,942
When paired with Gemma 4, EmbeddingGemma 2 powers efficient, on-device retrieval augmented generation (RAG) with a lower memory footprint. EmbeddingGemma 2 retrieves local files, and Gemma 4 reasons over them for grounded answers with complete privacy.
7
6
108
13,112
EmbeddingGemma 2 is optimized to run locally on mobile and desktop. This means you can find a specific video clip from a voice memo, or search through hours of audio with a simple text query, all without an internet connection.
13
14
145
23,560
We’re releasing EmbeddingGemma 2, our first natively multimodal open model engineered for on-device embeddings. Built on the Gemma 4 architecture and released under an Apache 2.0 license, it goes beyond text to unify images, video, audio, and code in a single embedding space.
120
262
2,638
1,035,021
Meet Nano Banana 2.1, our latest image generation and editing model 🍌 This upgraded version outperforms our previous models across the board, with notable leaps in visual design, mask-based editing, and subject consistency to help you create more natural-looking images.
309
635
5,381
1,186,628