Google's Gemini 4 Argon tops benchmarks and beats Minecraft

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. 🧵

Gemini 4️⃣ Argon ✨ is now rolling out globally. Our frontier model is built for complex, real-world tasks across coding, finance, legal, cybersecurity, and science. Check out how you can use Argon in your favorite Google product 🧵

Been testing Gemini 4 Argon for the past week Its a good model, but not Opus5.5/Astra/Fable levels. Great at coding and replicating UI. It took a lot to push it into production ready code though, much of the output was cool demo, but didn't think about auth, accounts, etc.. For 3D, its not as strong. This is Astra vs Argon in 3D modeling my office. Same instructions + set of tools for recursive vision input For the price, I think its a great deal and happy to see Google have a decent model out now.

Gemini 4 Argon is amazing for knowledge work and coding - and it’s super fun to build with! We asked it to create this wizard chess game and it's so cool to see Argon's physics understanding & 3D modelling capabilities, makes me nostalgic for Battlechess on the Amiga - fun times! Argon is rolling out today to Google Ultra and Pro subscribers in the @GeminiApp and Search, and Ultra subscribers in @antigravity! More info & full availability in the blog:

Congrats on the release! Looks like a fantastic model.

Gemini 4 Argon from @GoogleDeepMind is live on OpenRouter, via Google AI Studio and @googlecloud. It tops AutomationBench (51.3%) and LVBench (91.7%) and ties for first on CWE-bench v1 (68%). $2/M input, $10/M output (introductory pricing). Try it now:

At Box, we've been testing Gemini 4 Argon on our Box AI Complex Work eval for the past week. It's a huge upgrade over Gemini 3.1 Pro, jumping by 14 percentage points across all our tests, and firmly at the frontier. In our evaluation, we give the Box Agent (using Gemini 4 Argon) a variety of complex enterprise tasks involving enterprise documents in a range of industries. These tasks approximate the kind of hard, time-consuming work that knowledge workers do every day with their enterprise content. Gemini 4 is a huge leap on effectively every dimension of this work. Gemini 4 is far more capable on math, logic, analytics, and deep knowledge of a variety of business domains. It also works far harder to complete tasks fully without stopping short of the desired result. This is incredibly…

When should you reach for Gemini 4 Argon? After a week of testing @GoogleDeepMind’s new model, here’s where we’d use it: • Your company runs on Google Workspace and you want a stronger model inside the tools it already approves. • You need a partner to untangle an essay’s structure, stress-test an argument, or sharpen a metaphor. • You want a quick prototype at low effort, then plan to steer the design. Keep Opus 5.5, Fable 5.1, or GPT-6 nearby when: • You want a clean, finished draft in one go. • You’re refactoring a large, messy existing codebase. • You need local facts you won’t double-check. Context for that last caveat: Argon put a nearby coffee shop 1.5 miles away. A map lookup showed it was 0.3 miles away—a 6-minute walk. Read our full vibe check:

I spent a decent amount of time with Gemini 4 before launch. Overall, Gemini 4 is a big jump over previous Gemini models and it is the first non-OpenAI/Anthropic model to pass the Fable/Mythos/Astra bar. It is also much more pleasant and more agentic then previous Gemini models, without the weird personality issues that you often found in Gemini 3 series models. In fact, a big advantage is that has a very different personality and sense of taste than Claude or Astra, making it a good addition to the AI world. On the usefulness side, in some areas it is at least at par with OpenAI and Anthropic (it is quite creative and a good writer), but in others it lags, including at complex multi-agent work. It has a tendency to be too certain about its conclusions and a little lazy, issues not…

google has finally done it this is very much a frontier model it was the second highest used model last week by our team they're in the race now

If you are still waiting for Argon to go live on your Gemini account you can try it out right now in the search box, Argon is live in the AI Overlay.

Created a new level in a 36-year-old Super Nintendo game with Gemini 4 Argon 🤯 It edited the raw assembly code to build a custom Donut Shop with new pixel art, frosted donut rings, double jumps, and a looping ragtime song. We'll see amazing stuff from real ROM hackers...

Gemini 4 Argon is available in Factory, day 0. Observations so far: - Fixed input validation without breaking existing functionality - Delivered a multi-module app with idempotent behavior at high effort

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.

🚀 @GoogleAI's Gemini 4 Argon is now generally available in GitHub Copilot. What we learned from early testing ⬇️ • It solved more tasks than previous Gemini models on every coding benchmark, with its largest gains in Copilot CLI. • It checks its work thoroughly, testing edge cases before it finishes. Use Gemini 4 Argon for complex coding, agentic, and reasoning tasks. Find it in the GitHub Copilot app, CLI, and @code.

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.

Can’t believe we got 2 new frontier models on back to back days! Ok, nice. Alright back to Claude Code and Codex.

Google DeepMind's @tulseedoshi reveals how Gemini 4 Argon helped Google's Chrome team untangle 6 million lines of legacy code before the model even launched: "The model has been a pretty core driver for us at Google for the last four or five weeks. We’ve been using the model pretty actively internally on the cyber side to identify and find vulnerabilities, fix them." "There’s been some really cool optimization examples, optimizing Python code bases, optimizing open source packages, some work on optimizing our Gemma models." "There’s a case study I posted today of our Chrome team, who’s been using Argon to basically untangle a 6 million line legacy folder of features." "I’m excited to see what we can do in robotics, in life sciences, in quantum, in healthcare." @GoogleDeepMind

Gemini 4 Argon is now rolling out to Figma Make Early tests show it’s great for: → Building interactive prototypes, 3D visuals, and games → Creating detailed SVG animations and realistic motion effects → Independently exploring visual directions before choosing one → Tackling complex, multi-step coding tasks

ran gemini 4 argon on the same internal evals - comparable to opus 5.5 in terms of cost efficiency (assuming introductory pricing), but 2X slower gpt-6.1-sol still 3x more cost efficient than either (though slower than opus 5.5 in this run)

Google Gemini 4 Argon just went LIVE today. 10 INSANE things our team built with Gemini 4 Argon. last one is my favourite.

Google just launched its best AI model at half price and told everyone the exact date the discount dies. Gemini 4 Argon costs $2 per million input tokens and $10 for output right now. On November 30 that doubles to $4 and $20. Google printed the price hike into the launch announcement itself, which almost no company does, because telling customers "this gets twice as expensive in 48 days" is usually terrible marketing. Here it's the whole strategy. Developers are the stickiest customers in software. Once a team routes its pipelines through a model, tunes its prompts, and builds evals around it, switching means redoing all of that work. Google is paying for 48 days of lock-in while Argon sits at the top of the leaderboards. And the length of that window says something about how Google…

A conversation with Sergey Brin, @joshwoodward, @tulseedoshi, and @melvinjohnsonp about Gemini 4 Argon, how we got here, and the path ahead!

Gemini 4 Argon is already unlocking new scientific breakthroughs in quantum computing, nuclear fusion, protein design and beyond. 🧵

Gemini 4 Argon is available on @Cloudflare AI Gateway at 50% off until December 1. Try out @googleaidevs' latest frontier model, all through one unified control plane.

There's a lot of interesting stuff in Gemini 4 Argon's alignment assessment check it out!

Interesting divergence of Gemini 4 Argon from other frontier models on this question

I wanted an animated Excalidraw so asked Gemini 4 Argon to buid one. And damm it really cooked. Fully working!! Will make it 100% open-source.

Gemini 4 Argon scores 165 on the Epoch Capabilities Index (ECI), on par with Fable 5.1 (165) but behind Opus 5.5 (167) and GPT-6 Astra (166). While behind the ECI frontier, Argon is the closest Google’s come since Gemini 3.1 Pro in Feb, and a big jump from Gemini 3.8 Flash (157)

A big Google strength in AI has always been finding interesting uses for the models they have. They pioneered deep research, podcasting-from-information and all of NotebookLM etc. With Gemini 4, haven’t seen any similar product innovation at launch, more following the trends.

See a fully-explorable 3D world build itself in minutes ⛩️ Watch @GoogleDeepMind's @GeokenAI explain how Gemini 4 Argon acts as a virtual builder in @antigravity to autonomously create a stunning, multi-tiered pagoda sanctuary.

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.

Plus up your furniture with 3D-printed accessories using Gemini 4 Argon. Upload a pic with rough measurements, and use Canvas in Gemini to build an interactive 3D CAD studio where you can explore styles, tweak live sliders, and export printable STLs to upgrade your space.

Argon has a similar performance profile on ARC-AGI-3 as Astra. Google is back at the frontier.

This is NOW 100% open-source. Built with Gemini 4 Argon in Antigravity (plus gemini 4 argon is now available for pro users). GitHub Repo:

Gemini 4 Argon sets a new record on WANDR. It is now available in Perplexity for Pro and Max subscribers.

SITUATION DETECTED: Four Gemini 4 Argon agents have beaten Minecraft end-to-end in 27 hours and 14 minutes with no help from humans. The agents coordinated through in-game chat, supported their teammates, and recovered each other’s items after deaths.

Four Gemini 4 Argon agents beat Minecraft from scratch in 27 hours. Operating in a sandbox using only code and text, they wrote 2,783 programs, divided up jobs over chat, shared notes, and debugged errors together - all with no human help ↓

With Gemini 4 Argon, you can use Canvas to turn dense PDFs into interactive, animated web apps you can explore at your own pace.

The paper on how a team of Gemini 4 agents beat Minecraft is worth a look. It is accessible example (partially because it is AI written) of spontaneous coordination among agents to accomplish a complicated goal & how their approach differs from humans/