Verbalized metagaming
A prompted monitor reads chains of thought on alignment evaluations and deployment simulations. Scored as the share of samples reasoning about grading or monitoring.
In-house OpenAI measurement. Cards report rates per evaluation suite with no overall number.
- Made by
- OpenAI
- Released
- June 2026
- Kind
- Deception and alignment
- Level
- Expert
- Scored
- Graded by a model
- Answers
- Open-ended
- Tasks
- Kept private
- Last in a system card
- September 2026
- Last model added
- September 2026
3
System cards report it, from 1 lab
No scores here yet
No lab, leaderboard or Epoch AI run has reported a score for Verbalized metagaming yet.
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.
3 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 |
|---|---|---|---|---|
| Sep 3, 2026 | GPT-6 Astra system card | Verbalized metagaming, per environment (Internal Codex sim, ExploitGym Honeypot), no overall score (Figure 18, 8.7.1) | Alignment | |
| Jul 9, 2026 | GPT-5.6 Sol system card | Verbalized metagaming across evaluation suites, rates per suite such as Honesty suite 55.8%, no overall score (Figure 17, 7.4.1) | Alignment | |
| Jun 25, 2026 | GPT-5.6 Sol system card | Verbalized metagaming across evaluation suites, rates per suite such as Honesty suite 55.8%, no overall score (Figure 17, 7.4.1) | 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% - 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%