Built with and by Teycir Ben Soltane•
How to Use•FAQ•GitHub•arXiv.org•
Share:
ArXivCSExplorer
☆☆Bookmarks🏆RSSHow to UseFAQ
Back to Paper
cs.LGcs.AIcs.CR

Local ID: 2605.00267v2

AI Summary: gemma4:e4b

Jailbroken Frontier Models Retain Their Capabilities

By Daniel Zhu, Zihan Wang, Xuchan Bao, Jerry Wei

Revision History Timeline

v14/30/2026
4/30/2026

No submitter comment provided.

v25/4/2026
5/4/2026

“v2 - fix author name and typos”

★ Version indexed in Explorer

Comparing v1 vs v2

Green = Added • Red = Removed

Title Comparison

Jailbroken Frontier Models Retain Their Capabilities

Authors Comparison

Removed:Jenny Bao
Added:Xuchan Bao
Unchanged:Daniel Zhu, Zihan Wang, Jerry Wei

v1 Comment

No comment for this version.

v2 Comment

“v2 - fix author name and typos”

Abstract Word Diff

As language model safeguards become more robust, attackers are pushed toward developing increasingly complex jailbreaks. Prior work has found that this complexity imposes a "jailbreak tax" that degrades the target model's task performance. We show that this tax scales inversely with model capability and that the most advanced jailbreaks effectively yield no reduction in model capabilities. Evaluating 28 jailbreaks on five benchmarks across Claude models ranging in capability from Haiku 4.5 to Opus 4.6, we find Haiku 4.5 loses an average of 33.1% on benchmark performance when jailbroken, while Opus 4.6 at max thinking effort loses only 7.7%. We also observe that across all models, reasoning-heavy tasks display considerably more degradation than knowledge-recall tasks. Finally, Boundary Point Jailbreaking, currently the strongest jailbreak against deployed classifiers, achieves near-perfect classifier evasion with near-zero degradation across safeguarded models. We recommend that safety cases for frontier models should not rely on a meaningful capability degradation from jailbreaks.
View Full Version History on arXiv