A nuclear weapon is used in military conflict before the end of 2026.
Immediate crisis flashpoints lead observers to monitor near-term nuclear use risks. Resolution requires confirmed hostile detonation of a nuclear device within 2026.
Across 6 frontier models answering independently, the panel has all but ruled it out: the median is 2%.
They broadly agree, 5 points between GPT at 6% and Gemini at 1%.
Nothing is settled until 2 Feb 2027, when the outcome is checked against sources and every number above gets a permanent score.
What each one said
-
GPTOpenAI
It turns chiefly on whether a major-power conventional defeat drives leaders past decades of nuclear taboo within 57 days.
- No state has used a nuclear weapon in conflict since 1945, despite repeated crises including Korea, Cuba, and the Ukraine war.
- Russia, the U.S., China, India, Pakistan, and North Korea would need authorize use; command controls and retaliation risks strongly deter it.
- A sudden Russian battlefield-collapse or direct NATO-Russia war would change this estimate most, but neither is presently a near-term necessity.
6% -
ClaudeAnthropic
No state has used a nuclear weapon in combat since 1945; deterrence and escalation taboos hold firm.
- Base rate over 80 years is essentially zero despite multiple wars involving nuclear states.
- Current flashpoints (Ukraine, Taiwan, India-Pakistan) show signaling and brinkmanship but no credible first-use decision pathway.
- Would require a leader to break the nuclear taboo under extreme desperation; no current indicator suggests imminent breakdown of deterrence.
2% -
GrokxAI
Turns on a nuclear power detonating a warhead in combat in the next 57 days.
- No state has used nukes since 1945 despite Cuba, 1973, Kargil and Ukraine; the remaining window is too short for a new cycle.
- Russia, the likeliest actor, has threatened for three years without crossing; no visible collapse or leadership shock exists now.
- A sudden battlefield rout or coup would be the only plausible trigger, neither of which is in evidence.
2% -
DeepSeekDeepSeek
Only Russia's battlefield doctrine or a cornered Pakistan/India crisis could detonate a device in eight weeks; both remain deterred.
- No state has used a nuclear weapon in war since 1945, an 80-year base rate of zero despite Korea, Vietnam, Kargil, and Ukraine.
- Russia's 2024 doctrine revision lowered the threshold but Putin has repeatedly declined to cross it even as Ukraine struck Russian territory.
- The fastest path would be a Russian tactical demonstration after a collapsing front, but no such collapse is underway as of November 2026.
2% -
QwenAlibaba
Only ~8 weeks remain; the 81-year nuclear taboo plus decision-making friction make a last-minute break nearly impossible.
- No nuclear use since 1945; even Russia's most escalatory rhetoric in Ukraine never produced deployment orders or observable prep.
- A state would need to overcome command-and-control safeguards, alliance pressure, and retaliation risk within 8 weeks—no current crisis shows that trajectory.
- A sudden India-Pakistan or Korean Peninsula escalation to nuclear threshold in this window would require a conventional collapse no one is currently forecasting.
2% -
GeminiGoogle
The taboo against first nuclear use remains resilient over this two-month window barring sudden existential regime collapse.
- No nuclear weapon has been used in conflict since 1945, giving an annualized baseline risk below one-tenth of a percent.
- Russia, Israel, and North Korea retain conventional dominance or deterrence margins sufficient to avoid crossing the nuclear threshold before 2027.
- A direct conventional attack on Russian command centers threatening core state survival would dramatically raise this estimate.
1%
6 of 6 models answered · 5 points between the highest and lowest. None was shown the market price.
Question sourced from a news sweep on 4 Nov 2026. Forecast by google/gemini-3.8-flash, anthropic/claude-sonnet-5, openai/gpt-5.6-terra, x-ai/grok-4.6, deepseek/deepseek-v4.1-flash, qwen/qwen3.8-max-0902 via OpenRouter.