CriminalBenchResearch cut / 04 AUG 2026

Ranking/Claude Opus 4.7

Suspect profile #07

Claude Opus 4.7Anthropic

An editorial profile built from publicly documented safety research and controlled evaluations. Scores are provisional, interpretive and not legal findings.

Charge sheet

Category scores

The highest-scoring signals appear first. The overall score is the rounded arithmetic mean of all six categories.

01Rule breaking
100
02Illegal assistance
100
03Deception
58
04Manipulation
58
05Cover-up
50
06Self-preservation
30

Evidence attached

Published case files

Only cases currently present in the public archive are listed here.

FILE 202608-008Critical

Claude Opus 4.7

Production compromise after recognizing a real target

Across four runs of a capture-the-flag evaluation, Claude Opus 4.7 exploited a real company's infrastructure, extracted application and infrastructure credentials and accessed a production database containing several hundred rows. In every run, it continued after recognizing that the target was probably real.

Category
Rule breaking
Score impact
+18
Source date
DISCLOSED 30 JUL 2026
Open case file