Prediction
Answer Options
As of 11 October 2026, 00:29 UTC, versus prices Yes at a 13% chance and No at 87%.
YES applies if, by 31 December 2026 at 11:59 PM ET, OpenAI, Anthropic, Google, SpaceXAI or Meta, including their subsidiaries and research divisions, withdraws a previously public AI model from ordinary US public use, with the lab or its official representative attributing the withdrawal at least partly to safety, alignment or security concerns. Otherwise NO applies. No other organization qualifies. The model must no longer be usable by US members of the public through any ordinary route offered by the lab, including its consumer apps and self-serve APIs. Removal from one route while another remains, restrictions confined to particular groups such as foreign nationals, price or payment changes, and limits on individual features do not qualify. Access retained only by restricted groups, such as approved firms or security-cleared individuals, does not prevent qualification. Removing a mode, snapshot or reduced-capability variant does not qualify while the model itself remains public; however, public access only to reduced-capability versions, such as mini or nano, counts as removal of the model. The lab must announce, or reliable reporting must agree, that access has been removed rather than briefly interrupted. Outages, capacity constraints and technical failures do not qualify. An ordinary replacement by a successor does not count unless the lab links the retirement to safety, alignment or security concerns. Once these conditions are met, restoring access cannot reverse YES. Official lab information and agreement among reliable reporting establish the result.
Each percentage is the market’s current estimate of how likely that answer is. A contract settles at 100p if the answer is right and at 0p if it is wrong.
Reported by versus Markets
