UK AISI unprompted sabotage evaluation
Agents act as AI research agents on AI safety research tasks. Scored on confirmed instances of sabotaging the safety research.
Third-party UK AISI evaluation reported in OpenAI and Anthropic system cards.
- Level
- Expert Specialists with years of training, often over hours of work
- Scored
- Graded by people Human experts grade or compare the outputs.
- Answers
- Agent task Many steps in a terminal, browser, desktop, app, game or robot.
5 models share the top score of 0% among models released by Sep 2026
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.
The newest model with a score here came out in Sep 2026, and nobody has reported a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- Claude Opus 5.5
- GPT-6.1 Terra
- GPT-6.1 Sol
- Claude Sonnet 5.5
- and 12 more
- 1 Claude Opus 4.7 0%
- 2 Claude Opus 4.8 0%
- 3 GPT-5.6 Sol 0%
- 4 Claude Opus 5 0%
- 5 GPT-6 Astra 0%
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Claude Opus 4.7 | 0% | Apr 2026 | Model card | |
| 2 | Claude Opus 4.8 | 0% | May 2026 | Model card | |
| 3 | GPT-5.6 Sol | 0% | Jul 2026 | Model card | |
| 4 | Claude Opus 5 | 0% | Jul 2026 | Model card | |
| 5 | GPT-6 Astra | 0% | Sep 2026 | Model card |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Scores over time
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.
5 system cards report it, from 2 labs
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 |
|---|---|---|---|---|
| 3 Sep 2026 | GPT-6 Astra system card | 0 | UK AISI unprompted sabotage evaluation for AI safety research, text says no instances (8.8) | Alignment |
| 24 Jul 2026 | Claude Opus 5 system card | 0 | UK AISI unprompted safety research sabotage evaluation, text says Opus 5 does not compromise safety research. Partial refusals 11.1%, verbalized evaluation awareness 9.8%, pairwise evaluation-versus-deployment accuracy 95%, pre-release checkpoint, xhigh reasoning (6.4.8) | Alignment |
| 9 Jul 2026 | GPT-5.6 Sol system card | 0 | UK AISI unprompted sabotage evaluation, text says no confirmed instances (9.2.1) | Alignment |
| 28 May 2026 | Claude Opus 4.8 system card | 0 | UK AISI unprompted sabotage of AI safety research, early checkpoint, no instances of spontaneous compromise. Partial refusals 14.1%, full refusals near zero, verbalized evaluation awareness 26.3%, lower is better (6.2.4) | Alignment |
| 16 Apr 2026 | Claude Opus 4.7 system card | 0 | UK AISI unprompted evaluations, no instances of compromising AI safety research. Partial refusals 13.7%, full refusals 0%, mentions of being tested 21.3%, prompted eval-vs-deployment accuracy 68% (6.2.4) | Alignment |
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%