On August 1, 2026, OpenAI publicly disclosed “Astra,” describing an internal version of Astra as its “next major model.” You can read more about that here: https://openai.com/index/ten-advances-in-mathematics/. This market will resolve to "Yes" if OpenAI releases "Astra" or a model confirmed to be the same model referenced in the announcement described above, and that model is made available to the general public by the listed date (ET). Otherwise, this market will resolve to "No." A qualifying model must be named "Astra" (including variants or successor branding such as Astra 1, Astra 6, Astra X, GPT-6 Astra, or similar naming or rebranded versions) or be confirmed to be the same model referenced in the announcement by OpenAI or by a consensus of credible reporting. Products labeled as GPT-5.7, GPT-6, or similar will only count if they are confirmed to be the same model referenced in the announcement. For this market to resolve to "Yes," the qualifying model (as defined above) must be launched and publicly accessible, including via open beta or open rolling waitlist signups. A closed beta or any form of private access will not suffice. The release must be clearly defined and publicly announced by OpenAI as being accessible to the general public. The release must be either clearly defined and publicly announced by OpenAI as being accessible to the general public or otherwise made publicly accessible and explicitly labeled within the company’s official website. Labeling errors, placeholder text, or version names displayed on the website that do not correspond to a model that is actually accessible to the general public will not qualify. The primary resolution source for this market will be official information from OpenAI; however, a consensus of credible reporting will also be used.
OpenAI’s September 1 announcement that its forthcoming Astra model has reached the company’s first “critical” cybersecurity capability threshold under its Preparedness Framework is the main driver shaping trader views on release timing. The large language model demonstrated strong autonomous vulnerability discovery and exploit chaining on benchmarks like ExploitBench, prompting weeks of added safeguards and development pauses after an earlier unreleased model incident. OpenAI stated it plans to make Astra available “soon” but will initially restrict the most advanced cyber features to select testers and partners via its Daybreak Blue program, with broader access following defensive preparations. This measured rollout, alongside prior August previews of multi-agent math capabilities, highlights the tension between frontier performance gains and safety requirements that could influence exact launch windows.