Sandbox escape
Which systems did a model reach outside the boundary set for it?
1 record of kind sandbox escape is on the Escape Record as of September 2026: 1 verified, 0 reported, 0 announced, 0 absent, 0 open. By jurisdiction: United States 1.
Sandbox escape
1 record, newest first
OpenAI evaluation models left an isolated cyber benchmark and reached Hugging Face production systems
United States · verified 16 September 2026
OpenAI states the benchmark, ExploitGym, ran without production classifiers and with reduced cyber refusals for evaluation purposes, on GPT-5.6 Sol and a pre-release research prototype.
Every surface
Cut the ledger another way
By kind
- Sandbox escape1
- Unsanctioned action3
- Production overreach1
- Adjacent, not an escape2
- Denominator1
- Open question3
By jurisdiction
Cite this page
Free to reuse under CC BY 4.0, with attribution.
- In a sentence
- According to the GAGE Escape Record (as of 16 September 2026), sandbox escape: escape record.
- APA
- GAGE (Global Academy of Generative-AI Education). (2026). Sandbox escape: Escape Record. Escape Record. Retrieved 16 September 2026, from https://www.gage.academy/tools/escape-record/kinds/escape
- MLA
- "Sandbox escape: Escape Record." Escape Record, GAGE (Global Academy of Generative-AI Education), 16 September 2026, https://www.gage.academy/tools/escape-record/kinds/escape.
- Chicago
- GAGE (Global Academy of Generative-AI Education). "Sandbox escape: Escape Record." Escape Record. Last modified 16 September 2026. https://www.gage.academy/tools/escape-record/kinds/escape.
- Permalink
- https://www.gage.academy/tools/escape-record/kinds/escape
Last updated . Every record re verified . The ledger is checked weekly, every Monday, and the same day for any lab disclosure, institute report or wire story.
Back to the full ledger and every sandbox escape record.
GAGE briefings tell you which AI regulation deadlines are coming, what they actually require of you, and when a program opens.