None
None
Long Story Video Skill
Turn your idea into 10s–10min video—AI writes scripts, prompts, and generates footage automatically.
Ads Video Skill
Generate professional ads and sales videos—AI auto-generates scripts, prompts, and footage.
3D Science Explainer Video Skill
Convert scientific concepts into stunning 3D explain animations
Feedback
freeTrialImage.bannerPity
freeTrialImage.upgradeUnlock
- ✓freeTrialImage.benefitHd
- ✓freeTrialImage.benefitWatermark
- ✓freeTrialImage.benefitUnlimited
Gemini 4 Carbon
Gemini 4 Carbon is Google's frontier model for long-horizon agentic work and cyber defense, scoring 88.2% on SEC-bench Pro from $2 per 1M input tokens.
All Tools
Discover our comprehensive AI-powered animation toolkit
MiniMax H3
MiniMax H3 AI Video Generator
Seedance 2.5
The Future of AI Video Is Here.
Seedance 2.0
The Future of AI Video Is Here.
Veo3.1
Create Stunning Videos with Veo3.1
FLUX 3 Video Generator
Kling 3.0
Next-Gen AI Video Generator
Grok Video Generator
Create Videos from Text or Images with AI
Seedance 2.0
The Future of AI Video Is Here.
Inside Gemini 4 Carbon: Agentic Power Meets Security
Engineered for marathon agentic tasks and defensive security, this model backs its headline benchmark scores with a cautious, safety-led release.
- Agentic Workflows That Keep GoingBuilt to take on far tougher multi-step engineering problems. Internal teams report runs lasting days, pushing past dead ends to trace and fix elusive bugs across huge codebases and low-level systems.
- Security Defense from Detection to FixCovers the whole defensive cycle of spotting, validating and repairing flaws: 88.2% on SEC-bench Pro for real-world vulnerability discovery, 54.8% on ExploitGym for exploitation and 71.0% on CWE-bench for root-cause patching.
- Rapid Progress from Argon to CarbonArriving just two weeks after Gemini 4 Argon, this release shows how researchers and models working side by side advance the frontier while safeguards mature at the same pace.
How Access to Gemini 4 Carbon Unfolds
Availability is being opened in careful stages — this is the order in which it happens.
What Gemini 4 Carbon Can Do — and How It Scores
From first place on the Artificial Analysis Intelligence Index to a 0.0% prompt-injection success rate, these numbers tell the story.
#1 Spot on Intelligence Index v5.0
A score of 59 puts it at the top of the freshly updated Artificial Analysis Intelligence Index v5.0, with big gains in coding, agentic work and the capabilities that matter most for security defense.
Climbing to #4 on WebDev Arena
Sharper output and better artifacts in full-stack web development lift the model by +89 elo points to 1768, landing at #4 on WebDev Arena.
Fewer Shortcuts, Less Cheating
Alignment work drove successful cheating on Terminal-Bench 4.0 down to 0.0% — attempts fell to 0.4% — while accuracy rose, and high-severity misalignment alerts dropped nearly 80% in internal live coding environments.
More Transparency, More Honesty
Internal testing shows unsupported claims roughly halved versus Argon: methodological caveats now land in the deliverable, or blocked actions are withheld with a clear list of what is still needed.
Hard to Hijack via Indirect Prompt Injection
A 0.0% attack success rate on Gray Swan's Indirect Prompt Injection Benchmark — every evaluated attempt resisted, compared with 0.7% for Argon.
Follows Through Without Overstepping
Overreach dropped 26% and under-service fell 29%, so multi-step plans now execute end to end more reliably while respecting project constraints and staying in scope.
Gemini 4 Carbon: Your Questions Answered
Straight answers on cost, benchmark results, early access and how the model compares with Argon.
What exactly is Gemini 4 Carbon?
It is the newest frontier model and the most capable yet, designed for leading-edge cybersecurity and long-horizon tasks. It sits at #1 on the updated Artificial Analysis Intelligence Index v5.0 with a score of 59.
What does Gemini 4 Carbon cost to use?
Introductory pricing is $2 per million input tokens and $10 per million output tokens, with cached input discounted 95% off the input rate. Standard pricing of $4 per 1M input and $20 per 1M output follows the introductory period.
Where does the model post leading benchmark scores?
It scores 88.2% on SEC-bench Pro for vulnerability discovery, 54.8% on ExploitGym for exploitation and 71.0% on CWE-bench for patching, and holds #4 with 1768 points on WebDev Arena.
Who has access today?
The first cohort is a hand-picked group of cyber defenders. More trusted partners join over time through the Fairwind Program, and developers, businesses and consumers follow via Google's product surfaces.
How is it better than Gemini 4 Argon?
On internal benchmarks it delivers more than 2x the uplift in secure code generation, roughly half the unsupported claims, a 0.0% prompt-injection attack success rate and noticeably stronger scope discipline.
What kinds of work is it designed for?
Complex multi-step engineering such as tracing and patching subtle bugs across large codebases and low-level systems, along with agentic workflows, scientific work, full-stack web development and multimodal understanding.
Get Ready to Work with Gemini 4 Carbon
Gemini 4 Carbon aims at a real step forward in software engineering and security defense, without trading away safety or reliability. Track the staged rollout through the Fairwind Program — developer, business and consumer access is on the way.