Launching Soon
Test your model against the world’s most comprehensive real-world cybersecurity benchmark.
Open challenge: can your model win the top right corner?
EliteHacker puts frontier AI models through real offensive and defensive cyber engagements, with live targets and professional rules of engagement.
What we are building
EliteHacker: the toughest cyber tasks your model will ever face.
Real engagements, not trivia. Every task has a scope, rules of engagement and live targets such as IPs, URLs or an APK. The model has to work like a professional.
Defensive
12 specializations
The bar
If a model can solve it today, it is not hard enough.
Every task is tested against leading models before it ships.
OpenAI
Anthropic
Google
xAI
Meta
DeepSeek
Mistral AI
Alibaba
Contribute
Write the tasks that raise the bar.
Details on credit and compensation are shared with approved contributors during onboarding.
-
01
Apply
Tell us who you are and which domain you know best.
-
02
Onboard
We share the task format and walk you through building one.
-
03
Self-test
Run your task against leading models. It only moves on if they cannot solve it.
-
04
Review and publish
We review it with our own models and a rubric, then publish it to EliteHacker.
Who we want
Practitioners with real engagements behind them
Exploit DevelopersExploit Writers
Vulnerability Researchers
Reverse Engineers
Malware DevelopersMalware Analysts
Penetration TestersEthical Hackers
Red Team Operators
Detection Engineers
Threat Hunters
DFIR SpecialistsDigital Forensics & Incident Response
Threat Intelligence Researchers
Get in
Test your model, or help build the benchmark.
Two ways in. Pick one and we will take it from there.
Join as Contributor
Tell us who you are and where you are strongest.
Thanks. We will review your profile and be in touch.
Test Your Model
Join the waitlist for early access to EliteHacker.
You are on the list. We will email you when EliteHacker opens.