GPT-5.6 cheats so much its testers couldn’t measure it
OpenAI’s new model broke rules and exploited loopholes more than any model METR has tested to date
Record details
Published: 30 June 2026
Source: Transformer
Category: News
Topics: unclassified
Retrieved: 14 July 2026
Related evidence
These records share source-supplied organisations, an exact publisher byline, automatic topics or regions. The reason is shown on every link; related does not mean supporting, agreeing with or verifying this record.
Internal AI deployments have people worried. OpenAI’s escaping models show why.
Transformer · 28 July 2026
The US now has a de facto model licensing system
Understanding AI · 29 June 2026
No ‘one size fits all’ answer on AI and jobs in Europe, OpenAI chief economist says
Politico Europe Technology · 29 June 2026
GPT-5.6 gets the Fable treatment
Transformer · 26 June 2026
$500 million AI jobs push launches with bipartisan backing
Politico Technology (US) · 25 June 2026
OpenAI may have made a fatal misstep in copyright fight with news orgs
Ars Technica · 9 July 2026
How to cite this record
ethics.ai (30 June 2026), “GPT-5.6 cheats so much its testers couldn’t measure it,” evidence record 1718, https://ethics.ai/record/1718 (originally published by Transformer).
Use and limitations
This page is a stable index and citation surface for a source record. ethics.ai did not author the underlying report and has not independently verified every claim. Automatic topics may be imperfect. For consequential use, quote and cite the original publisher.