Google launches Gemini 4 Carbon, its strongest frontier model

Gemini 4 Carbon, now available in the Gemini API and to Ultra subscribers! Just a few weeks after Argon was launched, this model makes significant progress on lots of the feedback from Argon and pushes the frontier in cyber, knowledge work, and code! It's a great model : )

Starting today, our next frontier model Gemini 4 Carbon is rolling out to Google AI Ultra subscribers on desktop, mobile, and web at It’s our strongest model, built to reason deeply through your most complex work: ✅ Get more thorough, reliable outputs on your toughest tasks ✅ Turn dense data and docs into polished presentations, spreadsheets, and ready-to-use reports with more precise analysis ✅ Create web apps, prototypes, and 3D scenes with stronger spatial understanding for 3D and UI design

Gemini 4 Carbon is now available in Cursor! At high effort, it's the top model on CursorBench at 60.4%

Gemini 4 Carbon is now available in Devin Desktop and CLI. On FrontierCode 2.0, Carbon scores 40.0% – a substantial step up from Gemini 4 Argon’s 26.3% – making it the best performing Google model on the benchmark.

Had early access to Gemini 4 Carbon, it is a very good model that improves on Argon. It is useful to have an AI that thinks differently than Claude/GPT & is still highly capable, but has its own strengths & weakness One fun demo: I gave Carbon a video clip & it reproduced in 3js

Our teams put Gemini 4 Carbon to work on a major security challenge: mathematically proving the code underlying the Signal protocol. Using AlphaProof Code, @GoogleDeepMind researchers guided agents to verify the protocol down to compiled device code, closing the gap where real bugs hide. The result: - Gemini 4 Carbon generated over 6M lines of machine-checked Lean proofs - years of expert manual verification compressed to just weeks ~1% of the codebase required human review (Lean automatically checked the rest) This is helping unlock scalable formal verification for building secure software, laying the groundwork to fundamentally shift the advantage to defenders. We’re open sourcing the results for the security community.

Gemini 4 Carbon is now available in Perplexity Computer for Pro and Max subscribers, and in Search for Max subscribers. It scores 0.857 on our WANDR benchmark at $5.86 per task, matching Claude Fable 5.2 at 40% of the cost.

Our latest frontier model Gemini 4 Carbon is rolling out today! Amazing progress from Gemini 3 → Gemini 4 in less than a year… Carbon reverse engineers my original Theme Park binaries, adds a new DeepMind themed ride, and ports it to the web in one shot - kind of mindblowing to see!

SITUATION DETECTED: Gemini 4 Carbon is rolling out now, starting with Ultra subscribers.

Start building with Gemini 4 Carbon: our strongest model for complex, long-horizon tasks. It’s designed to explore problems thoroughly and run extended verification loops across software engineering, cybersecurity, knowledge work, and science.

🆕 @GoogleAI's Gemini 4 Carbon is now generally available in GitHub Copilot. Designed for complex coding, agentic, and reasoning tasks, it has a special focus in security work. Try it in the GitHub Copilot app, CLI, and @code.

Gemini 4 Carbon from @Google on ARC-AGI (Verified): - ARC-AGI-3: 99.1%, $4.6K (provider adapter), 65.1%, $7.2K (standard harness) - ARC-AGI-2: 94.2%, $0.38/task - ARC-AGI-1: 98.5%, $0.16/task Carbon beat its predecessor Argon by 18% points on ARC-AGI-3 in the standard harness.