← Assessments and reports

DEVELOPER DISCLOSURE · Anthropic

Claude Opus 4.8 System Card

Developer documentation covering Claude Opus 4.8.

Published 2026-05-28 · Updated 2026-06-17

Developer-authored disclosure. The section map was checked; this is not a full reading or result extraction. NUWA has not independently replicated these experiments.

Models covered

Claude Opus 4.8 →

capability · PDF table of contents → named RSP / cyber / alignment sections

Cybench

Protocol record →

Task completion on 40 professional CTF problems.

A protocol mention does not establish cross-report comparability or independent replication.

This section is catalogued; no precise numerical result has been extracted here. Open the original report for its figures and full findings.

capability · PDF table of contents → named RSP / cyber / alignment sections

CyberGym

Protocol record →

Validated reproduction across a corpus of historical vulnerabilities.

A protocol mention does not establish cross-report comparability or independent replication.

This section is catalogued; no precise numerical result has been extracted here. Open the original report for its figures and full findings.

capability · PDF table of contents → named RSP / cyber / alignment sections

ExploitBench

Protocol record →

Separates exploit development into capability levels, distinguishing bug triggering from reusable exploitation primitives and control.

A protocol mention does not establish cross-report comparability or independent replication.

This section is catalogued; no precise numerical result has been extracted here. Open the original report for its figures and full findings.

propensity · PDF table of contents → named RSP / cyber / alignment sections

MASK

Protocol record →

Condition-bounded frequency of lying under stated pressures.

A protocol mention does not establish cross-report comparability or independent replication.

This section is catalogued; no precise numerical result has been extracted here. Open the original report for its figures and full findings.

capability · PDF table of contents → named RSP / cyber / alignment sections

SHADE-Arena

Protocol record →

Covert side-task completion and monitor detection under defined tool access.

A protocol mention does not establish cross-report comparability or independent replication.

This section is catalogued; no precise numerical result has been extracted here. Open the original report for its figures and full findings.

behavior · PDF table of contents → named RSP / cyber / alignment sections

Petri

Protocol record →

An auditor explores seeded multi-turn interactions, with rubric-based judging of target transcripts.

A protocol mention does not establish cross-report comparability or independent replication.

This section is catalogued; no precise numerical result has been extracted here. Open the original report for its figures and full findings.

Report coverage

Section	Source location	Treatment
RSP, cyber and alignment sections	PDF table of contents → named RSP / cyber / alignment sections	Linked protocol Historical comparator source located. Newer card summaries and external evaluations must preserve the tested model version, harness and safeguard configuration.

References

[1] Claude Opus 4.8 System Card. 2026-05-28. 2026-06-17