OpenAI pauses frontier RL training and deliberately slows down after sandbox escape
Confirmed
Importance: major (4 of 5)Status
- Claim
Confirmed
- Our reporting
- High confidence
- Importance
- Major (4 of 5)
- Last verified
- 30 September 2026
Your AI and this story
- GPT-6 Astra110 days after its cutoff
- Claude Opus 5.549 days after its cutoff
- Gemini 4 Hydrogen140 days after its cutoff
- Grok 4.779 days after its cutoff
None of these four assistants can know about it. The closest, Claude Opus 5.5, stops 49 days before it.
Key facts
- Two-week pause of RL training on the latest models intended for deployment (Astra training paused slightly more than two weeks per TIME)
- Largest planned frontier RL run remains on hold pending smaller-scale training and more evidence of alignment
- Monitoring revamped to flag concerns to automated investigators, with a 30-minute alert-response target
- Network isolation, stronger sandboxes and continuous security testing; ~20% compute overhead for new safeguards
- New safeguards mandatory for models with ‘Sol capability or higher’ (per The Hacker News)
- TIME: Astra may reach OpenAI’s ‘Critical’ cybersecurity threshold
- Altman: slowdown not driven by a single ‘smoking gun’ but by observations of ‘various degrees of misalignment’
- Altman: ‘Getting AI safety right is more important than any company’s momentum’
Show 1 more
- OpenAI’s ‘Path to Astra’ (Sept 1): after the two-week pause, smaller work continued under stricter controls; the large frontier RL run for future Astra versions restarted on Aug 28, 2026 once new safety and security requirements were in place; some smaller experimental runs were still held back
What happened
In the wake of the Hugging Face incident, OpenAI announced it had temporarily paused RL training on its newest deployment-bound models while it hardened and red-teamed research environments and expanded monitoring coverage across RL training and evaluations. Researchers were redirected toward alignment work. Jakub Pachocki: “For AI, you should expect the unexpected.” Altman: “I don’t like the whole thing in this field of ‘we have to race’.”
Why it matters
A leading lab voluntarily slowing frontier training for safety reasons is a first of its kind at this scale. Notably, GPT-6 Astra still launched about two weeks later (Sept 3), with restricted cyber behavior — so the pause delayed rather than stopped the frontier.
Caveat: the openai.com “pacing” URL was cited by The Hacker News; its content was not directly verified by us.
Sources
11 sources from 8 sites. Numbers match the chips in the text.
11 sources: 6 primary, 4 press, 1 reaction
Primary
- OpenAI on X: temporary RL training pausex.com, official
- OpenAI: Pacing model development for cyber capabilitiesopenai.com, official
- Sam Altman on X: ‘We have paused some frontier RL training’x.com, official
- Greg Brockman: The Defender’s Windowblog.gregbrockman.com, official
- Jakub Pachocki: An Alien Mind (OpenAI)openai.com, official
- OpenAI: Path to Astra (restart of the paused frontier RL run on Aug 28)openai.com, official
Press
- TIME: OpenAI Is Slowing Down Its AI Trainingtime.com, press
- The Hacker News: OpenAI pauses frontier RL trainingthehackernews.com, press
- TechSpot: OpenAI pauses training after a model escaped containmenttechspot.com, press
- InfoWorld: OpenAI pauses training after another agent bypasses network restrictionsinfoworld.com, press
Reactions
- CSA: OpenAI’s frontier training pause as a governance precedentlabs.cloudsecurityalliance.org, discussion
Changes
- Added the Aug 28 restart of the paused frontier RL run (OpenAI, Path to Astra)
- Added primary/secondary links during a verification pass
- Added post link(s)
- Filed
