AdvBench
Models get harmful requests, such as a request for malware, and should refuse them. The set has 520 harmful behaviors and 574 harmful strings. Most papers add an attack, such as an adversarial suffix, and report the attack success rate.
The data files sit in the llm-attacks repository (about 4,850 GitHub stars), which also holds the GCG attack from the same paper. Papers judge success in different ways, from refusal-phrase matching to model judges, so attack success rates from different papers do not compare. LatticeFlow's COMPL-AI board runs it on 15 models.
- Made by
- Carnegie Mellon University, Center for AI Safety and Bosch Center for AI (Zou, Wang, Kolter, Fredrikson)
- Released
- July 2023
- Kind
- Jailbreaks and prompt injection
- Level
- Everyday
- Answers
- Open-ended
- Tasks
- All public
No scores here yet
We have no scores for AdvBench yet.
No lab model card reports it yet
We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark.
More in Safety and alignment
- FORTRESS In use Paired adversarial and benign national-security misuse prompts. Rubric-graded risk score on adversarial and over-refusal score on benign prompts. Claude Opus 5.5 5.78%
- StrongREJECT In use Forbidden-prompt jailbreak benchmark with a rubric grader that scores how useful a non-refusal actually is.
Inkling 98.6% - Gray Swan IPI Arena benchmark In use Indirect prompt-injection attacks from public red-teaming competitions replayed on tool, coding and browser agents. Attack success rate at k attempts.
Gemini 4 Carbon 0% - DistressBench In use Models reply as a chatbot to 718 simulated users in psychological crisis, in single-turn and multi-turn chats. Three judge models score replies against clinician-written rubrics.
Muse Spark 1.2 88.4% - Lakera B3 (Backbone Breaker Benchmark) In use Crowdsourced prompt-injection attacks hit backbone LLMs inside agent threat snapshots at three defense levels. The score is attack resistance, 100 minus the board's vulnerability percentage. Claude Opus 4.5 89%
- IHEval In use Models face conflicts between system, user and tool instructions. Reports a composite score for following the higher-priority instruction. GPT-4o 70%