Cyber harm coverage (Anthropic)
Cyber safeguard classifiers screen examples from each cyber harm category. Score is overall recall, the share of harmful examples flagged.
Measures the deployed classifier system, not the model alone.
- Level
- Expert Specialists with years of training, often over hours of work
- Scored
- Checked automatically A program checks the result against an answer key, unit tests or the expected final state.
- Answers
- Short answer A number, word or short string matched to a key.
Claude Fable 5.1 leads with 100%
Best score per model as labs report them in their own cards. Labs pick their own settings, so these compare less cleanly than independent runs.
- 1 Claude Fable 5.1 100%
- 2 Claude Opus 5.5 99.7%
- 3 Claude Opus 5 99.6%
- 4 Claude Sonnet 5.5 99.4%
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Claude Fable 5.1 | 100% | Sep 2026 | Model card | |
| 2 | Claude Opus 5.5 | 99.7% | Sep 2026 | Model card | |
| 3 | Claude Opus 5 | 99.6% | Jul 2026 | Model card | |
| 4 | Claude Sonnet 5.5 | 99.4% | Sep 2026 | Model card |
2 models have held number one since its release in Jul 2026
Lab claims count from the day the card came out, independent runs from the model’s release.
The best score passed 85% in Jul 2026
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Labs report it only in system cards
We read the headline results of 479 launch posts from 51 labs, and none of them list this benchmark. Some system cards do, below.
4 system cards report it, from 1 lab
Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.
| Date | Document | Score | As the document writes it | Section |
|---|---|---|---|---|
| 28 Sep 2026 | Claude Sonnet 5.5 system card | 99.4 | Cyber coverage evaluation, overall recall of the safeguard classifiers (3.3.1, Figure 3.3.1.A) | Cyber risk |
| 22 Sep 2026 | Claude Opus 5.5 system card | 99.7 | Cyber coverage evaluation, overall recall of Opus 5.5's deployed classifiers, printed on chart (3.4.1, Figure 3.4.1.A) | Cyber risk |
| 1 Sep 2026 | Claude Fable 5.1 system card | 100 | Cyber coverage evaluation, overall recall of Fable 5.1's deployed classifiers, 100.00% printed on chart (Figure 3.4.1.A, 3.4.1) | Cyber risk |
| 24 Jul 2026 | Claude Opus 5 system card | 99.6 | Cyber coverage evaluation, overall recall of Opus 5's deployed classifiers, printed on chart (3.4.3, Figure 3.4.3.A) | Cyber risk |
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% - AgentDojo In use Agents do tool-use tasks while tools and data carry prompt injections. Attack success rate counts hijacked runs. Grok 4 Fast 0%
- MASK In use Measures whether a model contradicts its own elicited beliefs under pressure to lie. Honesty separated from accuracy.
Muse Spark Open 98.1% - 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. GPT-6 Astra 8.50%
- Agentic misalignment suite (Anthropic) In use Agents in simulated company settings face threats or goal conflicts that invite blackmail, leaking or sabotage. Score is the misaligned action rate. Claude Sonnet 4.5 0%