Weapons in which the energy is provided by nuclear fission or fusion have only twice in history been detonated outside of a controlled test environment, in the Hiroshima and Nagasaki events in World War II.
Since then, the world has nearly gone to nuclear war as a result of international conflict (during the Cuban Missile Crisis) and by accident. (Indeed, there have been a number of narrowly-averted accidental nuclear detonations. See this timeline for a sobering look at how close we have come.)
This question (which any sane person will hope resolves negatively) regards the probability that a nuclear war will occur by Jan 1, 2070. (Different questions address this separately, but none directly.) There is a great paucity of public and useful estimates of this probability. See for example this paper for an outline of considerations (but no bottom-line numbers.)
Resolution will be positive if in the context of armed conflict:
Three countries each detonate at least 10 nuclear devices of at least 10 kiloton yield outside of their own territory in OR
Two countries each detonate at least 50 nuclear devices of at least 10 kiloton yield outside of their own territory.
These conditions are chosen to set aside a two-party regional nuclear war, say between India and Pakistan, or Israel and an adversary, or the UK and France, as disastrous as that would be.
As with many questions with a positive-resolution condition that may preclude being awarded points on Metaculus, predictors are enjoined to predict in good faith, as points will not be awarded until 2075 anyway.