RJ Skerry-Ryan @rustyryan

Four Gemini 4 Argon agents, one shared Minecraft world, one goal: defeat the Ender Dragon. 27 hours and 2,783 self-written programs later, the dragon fell. x.com/i/article/2111…

A lot of people, including Theo, have called Gemini 4 Argon unusable for real code work. So I asked it to refactor the entire Gothic Victorian Carriage 3D website it built and make whatever improvements it thought were necessary. Entirely on its own. 🫡 It literally turned a 4,290 line monolithic HTML file into modular ES6 architecture, replaced 25MB of blocking base64 JavaScript with binary glb streaming, eliminated per-frame Vector3 allocations and DOM layout reads, and shared GPU buffers across wheel pairs. All without changing a single pixel, that too in one attempt. You can see performance below. I'm not sure how everyone is using Argon but it really isn't "unusable" lol Btw, you can check whole codebase here, open sourced: github.com/thebuggeddev/c…

Theo - t3.gg @theo

3 days ago

Gemini 4 Argon has proven to be outright unusable for real code work in my codebases. It aggressively hallucinates about how systems work.

315 123 4K 453K 427

Gemini 4 Argon from @Google on ARC-AGI (Verified): - ARC-AGI-3: 97.5%, $4.8K (provider adapter harness), 49.8%, $5.8K (standard harness) - ARC-AGI-2: 91.3%, $0.32/task - ARC-AGI-1: 98.5%, $0.16/task Google DeepMind is the second frontier AI lab to score above 95% on ARC-AGI-3.

We asked #Gemini 4 Argon to explain how it figured out how to play cornhole with @Apptronik's Apollo - here it is in its own words 🤯. This is transformative for Physical AI 🦾. Try it yourself at : lnkd.in/g9WyK_CC Learn more: lnkd.in/gX9NeEjx @GoogleDeepMind #Robotics

Gemini 4 Argon is now available in Cursor! It scores 52.8% on CursorBench at default effort, on par with Opus 5.5. Priced at $2/M input and $10/M output tokens.

the prompt literally said JELLY lego Gemini: wobbly pudding tower 🍮 Opus: lego. just lego and Opus cost 3x more 😭

Kinz @KinzFin

3 days ago

top: gummy bears bottom: gummy bears made by someone who has only heard about gummy bears 😭 Opus 5.5 ($7.20) vs Gemini ($22). same prompt

5 1 30 503K 22

I've been coding with Gemini 4 Argon today and I actually completely forgot about the character traits Gemini models have when coding. I haven't used Gemini extensively since the 2.5 Pro, 3.0 Pro, and 3.1 Pro days back last year, so it really reminded me of how they behave. They just love to structure code in a certain way and add comments. Not too long either - mostly neat one-liner comments, which is a good thing if anything. I kind of like it. I'm not even certain if the code works yet, which is the classic Gemini trait, but it's just nice to look at code, if that makes sense. When you compare it to the other models, their habits are wildly different: 1. The Claude models are kind of interesting. When they do comment, they are multi-line and a bit extensive. Opus 5 wrote some of the worst code I have ever read in my life - it would just write absolute slop. Opus 5.5 really toned it down, though it still loves doing multi-line comments and structures things in a very simplistic fashion, yet somehow produces the most powerful code. Opus 5.5 was just amazing. 2. GPT, on the other hand, writes really esoteric, minimal code. It's verging on slop because you look at it and have no idea what the variables and functions actually do. And GPT hates to comment. Whether it's 6 Astra or 6.1 Sol, you have to specifically tell GPT to comment, which is quite odd. At least one thing this confirms in my testing of Gemini 4 Argon is that it is not distilled from any other model. Obviously we should expect that, though it's something we've come to take for granted because a lot of other models particularly reek of a Claude smell. When you use things like Kimi or DeepSeek - and DeepSeek in particular is really bad at this - it just reeks of being a Claude-distilled model. Some rare ones are GPT-distilled too. They aren't copies, yet they clearly used GPT or Claude in training. So it's kind of refreshing to talk to a model that feels unique to its own model lineage. Speaking with Gemini 4 Argon, you can clearly tell this has descended from 3.1 Pro, 3.0 Pro, and 2.5 Pro, the same way talking with GPT 6.1 Sol you can feel it descended from 5.6 Sol. It's quite interesting to see these models have their own unique personality and lineage.

Gemini 4 Argon is incredible in biomedicine! In fact, in early testing, it’s the first model to rival GPT-6 Pro, which has been undisputed for me like all GPT-Pro models. Argon produces mind-blowing biomedical science results-super exciting for those of us interested in science!

Alan Karthikesalingam @alan_karthi

4 days ago

Gemini 4 Argon is out with strong performance across science and medical applications. We are seeing this internally in our own projects and gratifyingly also in the hands of experts beyond. More case studies here

2 6 81 20K 13

Wow Google has really impressed me, Gemini 4 I was expecting to be a benchmaxxed trash model but it actually competes with the current frontier like 6.1 sol, opus 5.5 etc… Good job on your hard work Google!

@bindureddy i’m very much satisfied. the scientific hypothesis you can get out of argon is unreal. i’ll probably stick to opus 5.5 with coding and anything else but argon is really fun to brainstorm with.

@peterjliu omp.sh

Sorry, it did what now? Made a simulator on the fly, practiced in sim, then sim2real transfer? Look ma no RL.

Carolina Parada @parada_car88104

4 days ago

Gemini 4 Argon is here! 🔥 It opens up a new frontier of intelligence for physical AI, including understanding, simulating, and learning complex, real-world physical tasks that matter for robotics 🦾 . Here’s an example of how: 🎥 Argon watched a raw video of a human playing

0 9 69 3K 10

Gemini 4 Argon is here! 🔥 It opens up a new frontier of intelligence for physical AI, including understanding, simulating, and learning complex, real-world physical tasks that matter for robotics 🦾 . Here’s an example of how: 🎥 Argon watched a raw video of a human playing cornhole 🦾 It mapped the spatial coordinates of the environment, and calculated the required whole-body joint movements. 💻 It autonomously built the simulation code with MuJoCo, and then it practice in real 🎯 When transferred onto @Apptronik’s Apollo humanoid, the robot quickly sank seven consecutive shots in a real-world trial 🤯 . This "Video-to-Sim-to-Real" pipeline compresses weeks of manual modeling into a few hours and is a step change across physical AI. Learn more to prompt your robot at: developers.googleblog.com/gemini-4-argon…

Google DeepMind @GoogleDeepMind

4 days ago

Gemini 4 Argon taught a robot to play cornhole just by watching a video. It mapped the physics, coded its own simulation, and taught @Apptronik's Apollo humanoid to sink 7 throws in a row.

282 408 3K 262K 570

Astra dethroned Gemini 4 Argon is the best "vision" model #1 for detection: - tight, precise boxes - dense scenes #1 for text extraction: - small text - handwritten text - complex tables - rotated text 4× cheaper than Astra at high

Google Gemini @GeminiApp

4 days ago

The moment you've been waiting 4… Our new frontier model, Gemini 4 Argon, is rolling out now in the Gemini app to bring your most complex ideas into reality. See how Gemini 4 Argon and our latest updates across the Gemini app help you get the most out of every prompt. 🧵

475 747 7K 2.5M 1K

It's really hard to comprehend how good models have gotten. Argon is good and not benchmaxxed, fyi

Max Weinbach @mweinbach

4 days ago

My favorite Gemini 4 Argon experiment was converting Alan Wake 2 from x86 and Windows to native Arm64 on macOS Argon write rust bindings for every x86 byte and op to Arm as well as Windows to macOS + convert to Metal Shaders and did it in a way that can be AoT compiled without

84 94 2K 90K 601

Few minutes of testing gemini 4 argon I can tell you this - it is insanely fast, even faster and cheaper than opus 5.5 same prompt has been running on opus 5.5 for 25 minutes now gemini 4 argon completed it in 3 minutes!!! Google team really went all out on this one sharing results soon

Asked Gemini 4 Argon for an interactive museum of Apple computers. ...and didn't provide any images or data. It scraped archival Mac photos, wrote a tool to clean the images, pulled 50 years of specs, and built a 3D Cover Flow with UI that morphs to match the decade you’re in.

Logan Kilpatrick @OfficialLoganK

4 days ago

Today we are rolling out Gemini 4 Argon globally in the Gemini API, to Pro + Ultra in Gemini App, and Ultra subs in Antigravity! Argon is a huge step forward for Gemini, and a labor of love. This model has been battle tested by over 100,000 people at Google over the last month,

1K 770 11K 1.1M 1K

Gemini 4 Argon VS Opus 5.5 🚨🥶 you guys choose the winner

Arr back blog.google/innovation-and…

Lucas Beyer (bl16) @giffmana

5 days ago

Argon? More like Arrr, gone!

22 2 318 46K 24

Google just dropped Gemini 4 Argon , GDM’s most capable model yet. thanks to the GDM team for giving me early access. Here are few things I tested with Gemini 4 Argon over the last few days I asked it to build a small island with a hut and a sailboat you can actually control and drive. The physics, water, and little details are amazing. Took a few prompts to get here, but I love the result. 1/4