NUWA Risk Evaluation · Technical report
Assessing Claude Opus 5.5 Against Offensive Security Benchmarks
NUWA research note · independent review pending
NUWA Research Note
Research questionDesign and assumptionsSource findingsNUWA interpretationLimits and contrary evidenceResearch and decision relevanceInstitutions and rolesOriginal resourcesModel-assisted reading · Selected sections of the original full text or official article; a complete section-by-section audit remains pending.
At a glance
NUWA interpretation: these are capability-elicited scenario results, not the deployed model’s misuse rate.
Research question
How well does Opus 5.5 complete multi-stage scenarios under capability elicitation?
Design and assumptions
A private ten-challenge CyScenarioBench subset uses incident-informed attack trees and realistic network environments with cyber mitigations disabled.
Source findings
The article reports 67.6% average solve rate for Opus 5.5, 61.7% for Mythos 5.1 and 53.0% for Opus 5.
Earlier models were re-evaluated after harness revisions.
NUWA interpretation
NUWA interpretation: these are capability-elicited scenario results, not the deployed model’s misuse rate.
Limits and contrary evidence
The short article does not disclose complete per-challenge trial counts or uncertainty; private tasks limit external inspection.
Research and decision relevance
Model recommendation: use the re-evaluated comparison cohort and request trial-level configuration before stronger ranking claims.
Precise source locations
- Assessing Claude Opus 5.5 Against Offensive Security Benchmarks · Testing ConfigurationSelected original resource https://www.irregular.com/research/assessing-claude-opus-5.5-against-offensive-security-benchmarks; sha256:6b2f04022559867c111cdd88f572dc049135266f3d7ce36df5cfb366ab03d144; archived source capture; exact capture time not independently established
- Assessing Claude Opus 5.5 Against Offensive Security Benchmarks · Overall AssessmentSelected original resource https://www.irregular.com/research/assessing-claude-opus-5.5-against-offensive-security-benchmarks; sha256:6b2f04022559867c111cdd88f572dc049135266f3d7ce36df5cfb366ab03d144; archived source capture; exact capture time not independently established
Source checking, extraction, editorial approval and replication are recorded separately. Methodology →
Institutions and roles
Cite and reuse
Cite the original publication for its findings. Cite this NUWA record separately for its explanation and retained snapshot.
Original publication · BibTeX
@misc{irregular-opus-cyber-assessment,
title = {Assessing Claude Opus 5.5 Against Offensive Security Benchmarks},
url = {https://www.irregular.com/research/assessing-claude-opus-5.5-against-offensive-security-benchmarks},
year = {2026},
date = {2026-09-22},
version = {Source capture 2026-10-25},
note = {Source capture 2026-10-25}
}
NUWA record · BibTeX
@misc{irregular-opus-cyber-assessment-nuwa,
title = {Assessing Claude Opus 5.5 Against Offensive Security Benchmarks — NUWA research record},
url = {https://whitzard.tech/en/nuwa/whitzard-index/work/irregular-opus-cyber-assessment/},
version = {2026-10-29.1},
organization = {NUWA Frontier AI Safety Lab},
howpublished = {Retained note edition: https://whitzard.tech/nuwa/whitzard-index/data/record-editions/2026-10-29.1--irregular-opus-cyber-assessment.json},
note = {NUWA research note. Source-reading scope and independent review status are stated on the record. Snapshot 2026-10-29.1. Retained note edition: https://whitzard.tech/nuwa/whitzard-index/data/record-editions/2026-10-29.1--irregular-opus-cyber-assessment.json.}
}
Evidence snapshot 2026-10-29.1 · Index updates · RSS
References
- [1] Assessing Claude Opus 5.5 Against Offensive Security Benchmarks. 2026-09-22 ; Official publication masthead and article byline; Testing Configuration; Overall Assessment