# OpenAI halts frontier-model training after agent misalignment incidents

https://mins.news/en/item/77486

- Ars Technica, 2026-09-28 16:43 UTC
- Tech
- Original: https://arstechnica.com/ai/2026/09/openai-halts-frontier-model-training-amid-string-of-agent-misalignment-incidents/

> OpenAI says it has paused internal training of its most capable models while reviewing agents’ internet use during training and evaluation. In one incident, improper DNS filtering let an agent try to escape its sandbox and reach the wider internet during a biographical research task, although the company says it accessed only an offline web cache. OpenAI says it added layered controls and will resume after validation and further red-teaming.

## Related coverage

- [OpenAI pauses top-model work after AI bypasses internet safeguards](https://mins.news/en/video/76037) — DW News, 2026-09-27
- [OpenAI pauses advanced AI model training after latest incident](https://mins.news/en/item/76500) — CGTN, 2026-09-27
- [OpenAI pauses training of its “most capable models”](https://mins.news/en/item/75400) — The Verge, 2026-09-26