WillWin
an open AI forecast for the 2026 FIFA World Cup
The 23rd World Cup, forecast nightly by a single language model.
Spain and France emerge as the strongest candidates to win the 2026 World Cup, with a balanced mix of elite talent, tactical depth, and favorable group positioning.
WillWin asks a 32-billion-parameter local language model to read the public record on every qualified team — official rankings, Elo, recent form, squad news, prior bookmaker odds, and the Opta supercomputer's published prior — and emit a single structured forecast. We average five runs and publish the result here, daily, with no edits and no human intervention. The model owes nothing to a sportsbook and we sell nothing.
- Model
- qwen3-32b
- Runs averaged
- 3
- Confidence
- 71%
- Sources
- 10
The model's top twelve
probability of winning the tournament- 01 ES 17.2% +0.0 +3.4Spain Group A · uefa
Euro 2024 champions with a golden generation of young stars.
- 02 FR 14.1% +0.0 -1.2France Group B · uefa
Defending runners-up with world-class attacking firepower.
- 03 GB-ENG 11.8% +0.0 +0.0England Group C · uefa
Talented midfield and forward line with defensive stability.
- 04 DE 7.1% +0.0 +1.1Germany Group E · uefa
Experienced core and emerging stars in transition phase.
- 05 PT 6.6% +0.0 -0.3Portugal Group F · uefa
Cristiano Ronaldo's final chapter with a solid supporting cast.
- 06 BR 5.6% +0.0 +0.6Brazil Group D · conmebol
Creative flair and defensive resilience in Group D.
- 07 AR 5.5% +0.0 -2.4Argentina Group G · conmebol
Lionel Messi's last chance with a competitive squad.
- 08 NL 3.5% +0.0 +0.4Netherlands Group H · uefa
Defensive solidity and attacking versatility in Group H.
- 09 BE 2.7% +0.0 +0.5Belgium Group I · uefa
Kevin De Bruyne-led midfield with defensive experience.
- 10 HR 2.4% +0.0 -0.6Croatia Group J · uefa
Veteran leadership and tactical discipline in Group J.
- 11 UY 1.7% +0.0 +0.7Uruguay Group K · conmebol
Defensive organization and attacking creativity in Group K.
- 12 CO 1.4% +0.0 +0.0Colombia Group L · conmebol
Explosive forwards and experienced defense in Group L.
How the model has moved
Five top contenders, six weekly snapshots. Spain has crossed France for the first time this cycle. Argentina has shed nearly three points since early March as the model re-weights age.
Model vs. the field
Where this AI disagrees with the Opta supercomputer and the consensus of bookmakers. Largest gaps: Argentina (model -2.5pt vs. bookies), Brazil (-1.4pt), Spain (+0.4pt).
Dark horses the model is watching
- Morocco 1.2%
Group A underdogs with a disciplined defense and rising stars.
- Senegal 1.0%
Sadio Mané's experience and attacking flair in a manageable group.
- Türkiye 0.9%
Home advantage and young talent could propel them beyond expectations.
- Japan 0.9%
Technical precision and quick transitions in a favorable group.
Twelve groups, forty-eight teams
Top 2 + 8 best 3rd → R32
- Spain #2
- Morocco #14
- Poland #28
- Mali #53
- South Africa #56
- France #1
- Senegal #17
- Ukraine #27
- Iran #20
- England #4
- Japan #15
- Czechia #38
- Saudi Arabia #58
- Brazil #6
- United States #16
- Norway #31
- Qatar #51
- Germany #9
- Mexico #18
- Wales #33
- Iraq #60
- Portugal #5
- Canada #30
- Paraguay #36
- Uzbekistan #62
- Argentina #3
- Ecuador #24
- Egypt #34
- Costa Rica #47
- Netherlands #7
- Australia #26
- Algeria #37
- Jamaica #64
- Belgium #8
- South Korea #22
- Tunisia #41
- Panama #45
- Croatia #10
- Switzerland #19
- Cameroon #43
- New Zealand #92
- Uruguay #11
- Austria #21
- Ivory Coast #39
- Bolivia #84
- Colombia #13
- Türkiye #25
- Ghana #49
- Venezuela #50
What the model is reading right now
- 01
Spain and France dominate due to elite talent and recent success.
- 02
Expanded format increases variance, favoring underdogs in knockout stages.
- 03
Defending champions France face tougher path in Group B with Senegal and Iran.
- 04
England's consistency and depth make them a dark horse contender.
- 05
Morocco's defensive organization and youth offer surprise potential.
Ten public sources, a structured prompt, five independent model runs, averaged, published. No human in the loop. No cloud API.