A direct military conflict between two or more nuclear-armed superpowers begins before 2030.

Question

Six frontier AI models put this at 8% by 15 Sept 2031. Full reasoning from every model.

WillHappen.ai · Accepted Answer

8% — the median of 6 frontier models, resolving by 2031-09-15.

Answer

Gemini (Google): 9% — Mutual assured destruction continues to channel friction between the US, Russia, and China into proxy wars rather than direct combat. No direct clash between nuclear superpowers has occurred since the 1969 Sino-Soviet skirmish, establishing an extremely low base rate. A direct engagement before 2030 requires either a rapid Taiwan amphibious assault or direct NATO intervention in Ukraine. A formal US-China maritime encounter with casualties or explicit NATO kinetic strikes on Russian soil would shift odds upward.

Answer

Claude (Anthropic): 6% — Hinges on whether Russia-NATO or India-Pakistan/China friction crosses into direct state-on-state combat, not proxy war. No direct armed clash between nuclear-armed states has occurred since 1969 Sino-Soviet border clashes, showing strong historical avoidance despite tensions. Ukraine war, Taiwan tensions, and India-Pakistan skirmishes remain proxy or limited/non-nuclear-state conflicts, not direct nuclear-power-vs-nuclear-power war. Mutual deterrence and escalation-control incentives remain intact; a shift would require a major crisis like a NATO-Russia direct clash or China-India war escalating…

Answer

GPT (OpenAI): 19% — It turns on whether a Taiwan, NATO-Russia, or India-Pakistan crisis escapes leaders’ strong incentives to cap violence. Since 1945, nuclear powers have avoided sustained direct war despite Korea, Cuba, Kargil, and repeated NATO-Russia crises. Beijing, Moscow, Washington, Delhi and Islamabad must authorize overt attacks by 2030; nuclear deterrence and allies’ restraint raise the bar. A Chinese move against Taiwan or a lethal NATO-Russia exchange with treaty-triggering casualties would sharply raise this estimate.

Answer

Grok (xAI): 7% — Direct US-Russia or US-China shooting war, not proxies. NATO/Russia and US/China have avoided kinetic clashes for 80 years despite Cuba, Ukraine, Taiwan crises. MAD plus 3.3-year window leaves only accident or Taiwan invasion as plausible triggers, both low-base-rate. A Chinese amphibious assault on Taiwan or Russian strike on NATO territory would most shift this.

Answer

DeepSeek (DeepSeek): 12% — Whether a US-Russia or US-China exchange of fire crosses the direct-clash line in the next five years. The actors who could start it are Putin, Xi, and the US president; each has so far chosen proxy or gray-zone action over direct fire. The base rate for direct nuclear-power war since 1945 is roughly one episode (1969 Sino-Soviet border) in 80 years, so ~1-2% per year. Ukraine and Taiwan are the live triggers, but NATO and Beijing have repeatedly absorbed provocations without firing on each other's forces.

Answer

Qwen (Alibaba): 7% — Requires two nuclear-armed states to initiate direct conventional strikes on each other's forces, not proxies or gray-zone ops. Nuclear-armed states have fought zero direct wars against each other since 1945; the only near-misses (Cuba '62, Kargil '99) de-escalated fast. Russia-NATO remains the most plausible path, but both sides have consistently limited escalation to avoid Article 5 triggers through 2026. A sudden regime collapse or accidental engagement in the Baltics or South China Sea could shift this, but no such trigger is currently in motion.

A direct military conflict between two or more nuclear-armed superpowers begins before 2030.

Palestinian officials denied US visas attend the UN General Assembly in New York.

A bilateral or trilateral ceasefire between Russia and Ukraine is officially announced.

The US and Iran hold a formal senior-level diplomatic meeting by September 30, 2026.

The US announces the end or suspension of its naval blockade of Iran by December 31, 2026.