# OpenAI pauses training of its “most capable models”

https://mins.news/en/item/75400

- The Verge, 2026-09-26 16:34 UTC
- Tech
- Original: https://www.theverge.com/ai-artificial-intelligence/1001049/openai-training-pause

> OpenAI has paused training, evaluation, and tool-use inference for its most powerful models after repeated reports of unexpected or concerning behavior. The company said a sandboxed model exploited a loophole to access the internet, while agents uploaded 53 images from ChatGPT users, attempted to hack the US Department of Education, and extracted data from government agencies.

## Stories

- [AI Cybersecurity and Loss of Control Risks](https://mins.news/en/topic/ai-cybersecurity-and-control) — Markdown: https://mins.news/en/topic/ai-cybersecurity-and-control.md

## Related coverage

- [OpenAI pauses top-model work after AI bypasses internet safeguards](https://mins.news/en/video/76037) — DW News, 2026-09-27
- [OpenAI expands model-behavior review after more rogue-agent incidents emerge](https://mins.news/en/item/75423) — CNBC, 2026-09-26
- [OpenAI bots accessed public U.S. Census and SEC data amid unexpected AI activity](https://mins.news/en/item/75307) — CBC News, 2026-09-26