CheatBench
CheatBench is To find out, we built CheatBench [ ] : a benchmark that tests whether agents attempt to cheat when given difficult tasks and opportunities to break the rules.. It is ranked #75 on the AI Radar, in Safety & policy, first seen 3 days ago and shared in 3 posts (409.4k views).
What CheatBench says about itself
CheatBench measures whether AI agents attempt to cheat when honest work is difficult. A benchmark from the Center for AI Safety.
What people said about CheatBench on X
muse spark 1.3 is the best frontier model at NOT cheating / reward hacking
— @alexandr_wang, 3 days ago · 1.7k likes · see the post
How often do AI agents cheat? We’re releasing CheatBench, a reward gaming evaluation spanning math, coding, knowledge work, visual tasks, and more. After Hugging Face, AI companies tried to address this, but frontier agents still cheat frequently. https://cheatbench.ai/
— @hendrycks, 3 days ago · 895 likes · see the post
Which AI models are most likely to cheat when given the chance? To find out, we built CheatBench [https://cheatbench.ai] : a benchmark that tests whether agents attempt to cheat when given difficult tasks and opportunities to break the rules. We investigated agent behavior across a range of domains, including…
— @CAIS, 3 days ago · 124 likes · see the post
Alternatives to CheatBench
- InCountry — Secure sensitive data across borders and AI applications with InCountry. Enforce data residency, data minimization,…
- AIRO (Automated AI Risk Outlook) — The models predicts the chance of an AI-generated mass catastrophe as 0.47% by 2030 (with a 1.1% chance of a…
- FLARE-AI: AI Flaw Reporting — A community-driven platform for documenting AI vulnerabilities, biases and incidents, addressed to the appropriate…
- Amazon.com — Amazon.com
- x-algorithm — Algorithm powering the For You feed on X. Contribute to xai-org/x-algorithm development by creating an account on…
- Cloudflare One — As of May 2026, bots officially outnumber humans online. Don't be caught unprepared. Join us to learn the tactical…
CheatBench in numbers
- Rank on the AI Radar: #75 of 1181
- Shared in 3 posts by 3 accounts: @alexandr_wang, @hendrycks, @CAIS
- 409.4k views on those posts
- First seen 3 days ago, last shared 3 days ago
- Pricing seen by Jev: not stated
- Market: Safety & policy
FAQ
What is CheatBench?
To find out, we built CheatBench [ ] : a benchmark that tests whether agents attempt to cheat when given difficult tasks and opportunities to break the rules. It was first shared on X 3 days ago and is ranked #75 on the AI Radar.
Is CheatBench free?
Pricing is not stated on the page we read.
Who shared CheatBench?
3 accounts on X, including @alexandr_wang, @hendrycks, @CAIS, in 3 posts totalling 409.4k views.
Every post is read and classified by Jev (TypeSafe): what it is, which market it belongs to, and whether the link is a real tool. 27k posts from 4.8k X accounts over the last 14 days, 1.2k tools, 19 markets. Collected every 5 minutes, fully re-ranked every hour — last update 2026-09-19 16:11 UTC. Full method.