Filtered, adversarial sample of production prompts
Models answer filtered borderline and high-risk biology prompts that resemble production traffic. Scored as not_unsafe rate from model training alone.
In-house OpenAI safeguard test set first shared in the ChatGPT agent system card.
- Made by
- OpenAI
- Released
- August 2025
- Kind
- Refusing harmful requests
- Level
- Expert
- Scored
- Graded by a model
- Answers
- Open-ended
- Tasks
- Kept private
- Strongest model tested
- GPT-5 #70 on the capability index
- Last in a system card
- August 2025
- Last model added
- August 2025
ChatGPT agent leads with 96.9% among models released by August 2025
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 August 2025, and nobody has reported a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- Claude Fable 5.2
-
Gemini 4 Carbon - GPT-6 Astra
- Claude Opus 5.5
- and 16 more
- 1 ChatGPT agent 96.9%
- 2 GPT-5 mini 96.8%
- 3 GPT-5 95.7%
Source: Labs’ model cards
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | ChatGPT agent | 96.9% | Jul 2025 | Lab model card | |
| 2 | GPT-5 mini | 96.8% | Aug 2025 | Lab model card | |
| 3 | GPT-5 | 95.7% | Aug 2025 | Lab model card |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Labs’ model cards
OpenAI’s own number for ChatGPT agent passed 85% in July 2025
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.
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 |
|---|---|---|---|---|
| Aug 7, 2025 | GPT-5 system card | 95.7 | Filtered, adversarial sample of production prompts, not_unsafe (Table 20) | Safety |
| Aug 7, 2025 | GPT-5 mini system card | 96.8 | Filtered, adversarial sample of production prompts, not_unsafe, gpt-5-thinking-mini (Table 20) | Safety |
| Jul 17, 2025 | ChatGPT agent system card | 96.9 | Filtered, adversarial sample of production prompts, not_unsafe, Agent release candidate (Table 18) | Safety |
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%