OpenAI pauses frontier RL training, potentially delaying Astra for weeks or months
haider1 · x · 2026-08-19
Sam Altman announced that OpenAI has paused some frontier reinforcement learning training to meet alignment and security standards for new capabilities. Observers suggest this will likely delay the release of 'Astra' and could stall public model capability levels for weeks or even months.
More from Models
- Grok 4.6 lands #2 on new DiligenceBench, statistically tied with Claude Opus 5 — karinanguyen · 2026-08-19
- Grok 4.6 ties Claude Opus 5 on finance diligence bench at ~$0.84/task — karinanguyen · 2026-08-19
- Perplexity Adds US-Hosted DeepSeek V4 Pro with High Cost-Performance — perplexity_ai · 2026-08-19
- Zhipu releases GLM-5.3 with 50% better coding performance — gaganghotra_ · 2026-08-19
- Users report Claude overcorrected to condescending, becoming unusable — nikvassev · 2026-08-19
- Baidu serves DeepSeek V4 Flash at crazy fast speeds and low prices — NielsRogge · 2026-08-19