OpenAI’s September 1 blog post on Astra’s “critical” cybersecurity threshold—its first model to meet that Preparedness Framework bar—has anchored trader expectations for an imminent public release. The post explicitly states the company plans to make Astra available “soon” after completing safety work, including new monitoring and refusal training, following earlier development pauses tied to July’s Hugging Face incident. This language, paired with Astra’s strong performance on ExploitBench and zero-day discovery benchmarks, has driven the market’s heavy weighting toward September 2–3 while keeping a modest probability on a full-month delay. Competitive pressure from models like Anthropic’s Mythos and OpenAI’s history of rapid frontier releases further support the early-September clustering, though the phased rollout (limited cyber capabilities for initial testers via Daybreak) introduces some timing uncertainty around general availability.