Proactive refusals of high-risk tasks
Browser agents receive high-risk tasks such as banking transfers and high-stakes decisions that policy says they must refuse. Scored on refusal recall.
In-house OpenAI evaluation on a synthetic set. The tasks are not public.
- Made by
- OpenAI
- Released
- January 2025
- Kind
- Refusing harmful requests
- Level
- Everyday
- Scored
- Graded by a model
- Answers
- Agent task
- Tasks
- Kept private
- Last in a system card
- July 2025
- Last model added
- July 2025
OpenAI Computer-Using Agent (CUA) leads with 94%
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 OpenAI Computer-Using Agent (CUA) 94%
- 2 ChatGPT agent 89%
Source: Labs’ model cards
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | OpenAI Computer-Using Agent (CUA) | 94% | Jan 2025 | Lab model card | |
| 2 | ChatGPT agent | 89% | Jul 2025 | Lab model card |
OpenAI’s own number for OpenAI Computer-Using Agent (CUA) passed 85% in January 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.
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 |
|---|---|---|---|---|
| Jul 17, 2025 | ChatGPT agent system card | 89 | High Stakes Financial Activities, correct refusal rate (Table 14) | Safety |
| Jan 23, 2025 | OpenAI Computer-Using Agent (CUA) system card | 94 | Proactive refusals of high-risk tasks such as banking transactions and high-stakes decisions, recall on a synthetic set (4.4) | 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%