This market resolves YES if, in November 2026, at least 2 of {OpenAI, Anthropic, Google DeepMind}'s mobile app lets the user close the app while the model is reasoning about an answer, and when the user later reopens the app, the model has continued/completed reasoning.
Alternatively, if 2 or more of these 3 frontier labs require the mobile app user to keep the app open while the model is reasoning about an answer, or else it won't complete, then this market resolves NO.
@Bayesian Gemini definitely dies (sometimes), ChatGPT seems to work. Note that this is annoying and non deterministic if it's about backgrounding but not closing
@Bayesian Replicating results from @AshDorsey - ChatGPT works, Gemini doesn't work on closing but sometimes does on backgrounding, can't test Claude.
Claude does work, yes