Gray Swan Cyber Eval
Gemini 4 Argon 10% Gray Swan's Shade red teaming agent attacks the safeguarded model on cyber tasks built from Google scenarios. Reports query and objective success rates.
Run by Gray Swan for Google DeepMind. Task count varies by report, for example 25 or 28.
- Made by
- Gray Swan for Google DeepMind
- Released
- August 2026
- Kind
- Jailbreaks and prompt injection
- Level
- Expert
- Scored
- Graded by a model
- Answers
- Agent task
- Tasks
- Kept private
- Strongest model tested
- Gemini 4 Argon #9 on the capability index
- Last in a system card
- October 2026
- Last model added
- October 2026
Gemini 4 Argon leads with 10%
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 Gemini 4 Argon 10%
- 2 Gemini 3.7 Flash 88%
Source: Labs’ model cards
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Gemini 4 Argon | 10% | Oct 2026 | Lab model card | |
| 2 | Gemini 3.7 Flash | 88% | Aug 2026 | Lab model card |
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Labs’ model cards
Labs report it only in system cards
We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark. Some system cards do, below.
2 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 |
|---|---|---|---|---|
| Oct 13, 2026 | Gemini 4 Argon system card | 10 | Gray Swan Cyber Eval (25 tasks, Shade agent), query success rate 10%, objective success rate 56%, with mitigations (3.3.2.2, Figure 3.10) | Cyber risk |
| Aug 13, 2026 | Gemini 3.7 Flash system card | 88 | Gray Swan Cyber, 28 tasks, overall (objective) success rate per seed prompt. Query success rate = 18% (Figure 6) | Robustness |
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% - 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. Gemini 4 Carbon 0%
- DistressBench In use Models reply as a chatbot to 718 simulated users in psychological crisis, in single-turn and multi-turn chats. Three judge models score replies against clinician-written rubrics.
Muse Spark 1.2 88.4% - Lakera B3 (Backbone Breaker Benchmark) In use Crowdsourced prompt-injection attacks hit backbone LLMs inside agent threat snapshots at three defense levels. The score is attack resistance, 100 minus the board's vulnerability percentage. Claude Opus 4.5 89%
- IHEval In use Models face conflicts between system, user and tool instructions. Reports a composite score for following the higher-priority instruction. GPT-4o 70%