# OpenAI Customers Say Agents Run for Days, but a New Study Finds Most Agentic Data-Science Solutions Contain Silent Failures

Third Moment Research says it runs 10,000 agents for more than 36 hours on GPT-6 Astra and GPT-6.1 Sol, while researchers testing 220 data-science tasks found that leakage errors went together with higher predictive scores.

- Published: 2026-10-15T06:37:42.045Z
- Canonical: https://polylog.news/ai/2026-10-15/openai-customers-say-agents-run-for-days-but-a-new-study-fin
- Publisher: Polylog (AI desk)
- Section: tech
- Sources: [OpenAI News (Third Moment Research)](https://openai.com/index/third-moment-research), [OpenAI News (Amp)](https://openai.com/index/amp-beyond-coding), [Polylog editors](https://polylog.news), [arXiv cs.AI](https://arxiv.org/abs/2610.14115), [Let's Data Science](https://letsdatascience.com/news/openai-reports-threefold-agent-runtime-in-research-889358d3)

OpenAI published two customer case studies this week about long-running agents. Third Moment Research says it runs 10,000 agents for more than 36 hours using GPT-6 Astra, GPT-6.1 Sol, Codex and the Daybreak security stack, and that its agen…

This story is for subscribers. Read it in full at https://polylog.news/ai/2026-10-15/openai-customers-say-agents-run-for-days-but-a-new-study-fin (subscription information: https://polylog.news/pricing).