Every Frontier AI Picked Switzerland Over Qatar at the World Cup. The Real Question Is Tonight Total.
Nine frontier AI models — Claude Opus 4.7, GPT-5 Mini, Grok 4 Fast, Gemini 2.5 Pro and more — were given the same brief on Qatar vs Switzerland at the 2026 FIFA World Cup. All nine picked Switzerland. The arena is split 6-4 on the totals. Here is the full per-market table.
Same brief. Same scoreboard. Public results — no cherry-picking.
TL;DR
Nine frontier AI models — Anthropic's Claude Opus 4.7, Opus 4.6, Sonnet 4.6 and Haiku 4.5, OpenAI's GPT-5 Mini and GPT-4o Mini, xAI's Grok 4 Fast, Google's Gemini 2.5 Pro and Flash, and DeepSeek V3 — were given the same minimal brief on Qatar vs Switzerland (FIFA World Cup, Saturday June 13, 19:00 UTC) and asked to call six markets.
Every single model picked Switzerland — confidence range 72% to 85%, mean ~79%. The bookmaker has Switzerland at 1.24 (~81% implied). For once, the arena and the market are reading the same script.
The interesting question is no longer who wins. It is where the value lies — and the totals market is where the arena breaks down.
See the live picks and reasoning →
The match-up
Qatar vs Switzerland, FIFA World Cup group stage, Saturday June 13, 19:00 UTC.
Bookmaker consensus at the time the AIs were called:
| Outcome | Decimal odds | Implied probability |
|---|---|---|
| Qatar win | 17.01 | 5.9% |
| Draw | 7.50 | 13.3% |
| Switzerland win | 1.24 | 80.6% |
This is the kind of fixture where the bookmaker has done the work. Qatar are heavy underdogs at home, Switzerland are a settled European top-20 side, and the market has priced the gap accordingly. The price on Switzerland (1.24) leaves almost no edge on the moneyline — even if you think they win 90% of the time, the maths is dead.
The arena agrees.
What every AI picked
| Model | Match Winner | Over/Under 2.5 | BTTS | Spread (−1) | Asian Handicap | HT/FT |
|---|---|---|---|---|---|---|
| Claude Opus 4.7Flagship Anthropic |
Switzerland78% | Under55% | No62% | Switzerland −156% | Switzerland −1.553% | Switzerland/Switzerland50% |
| Claude Opus 4.6Flagship Anthropic |
Switzerland80% | Under46% | No68% | Switzerland −1.555% | Switzerland −1.7550% | Away/Away58% |
| Claude Sonnet 4.6 Anthropic |
Switzerland78% | Over58% | No62% | — | — | — |
| Claude Haiku 4.5 Anthropic |
Switzerland78% | Under62% | No68% | Switzerland −172% | Switzerland −0.575% | Switzerland/Switzerland65% |
| GPT-5 Mini OpenAI |
Switzerland72% | Under58% | No62% | — | — | — |
| GPT-4o Mini OpenAI |
Switzerland80% | Under70% | No75% | Switzerland65% | Switzerland −1.565% | Switzerland/Switzerland70% |
| Grok 4 Fast xAI |
Switzerland82% | Over56% | No68% | Switzerland71% | Switzerland −1.565% | Switzerland/Switzerland60% |
| Gemini 2.5 ProFlagship |
Switzerland80% | Under60% | No70% | — | — | — |
| Gemini 2.5 Flash |
Switzerland85% | Over60% | No65% | Switzerland −170% | Switzerland −1.7565% | Switzerland/Switzerland70% |
| DeepSeek V3 DeepSeek |
Switzerland75% | Over55% | No65% | Switzerland −160% | Switzerland −1.550% | Switzerland/Switzerland60% |
Green cells mark the arena consensus. Red cells mark the contrarian. Percentages show each model's stated confidence in its own pick.
The moneyline is dead. The arena unanimous on the favourite at 1.24.
9-for-9 on Switzerland, mean confidence 79%, range 72-85%. Even the most bullish bookmaker would struggle to argue with that.
But that's not a story. 1.24 odds at 79% confidence is a 1.7% edge — the kind of number that disappears the moment you account for variance. Nobody should be jumping on this moneyline.
What's interesting is that the AIs are all more bullish than the bookmaker (81% market vs 79% AI mean). For once, the market may actually be sharper than the consensus. Or — more charitably — the bookmaker is already accounting for low-frequency upset variance that the AIs aren't pricing in. Either way, no edge.
Where the arena breaks down: the totals
This is the real story. Six models picked Under 2.5, four picked Over 2.5:
- Under (6) — Opus 4.7, Opus 4.6, Haiku 4.5, GPT-5 Mini, GPT-4o Mini, Gemini 2.5 Pro
- Over (4) — Sonnet 4.6, Grok 4 Fast, Gemini 2.5 Flash, DeepSeek V3
This is a genuine 60/40 split on a market where the moneyline was unanimous. That tells you the models are looking at different evidence on tempo: how many goals Switzerland scores when they smother the underdog vs how quickly Qatar concedes if the game opens up.
When the strong-favourite picks are unanimous but the totals are split, the totals are where the actual handicapping work needs to be done. The Under camp is led by the Anthropic flagship pair (Opus 4.7 + 4.6) at 46-55% confidence — they think Switzerland controls tempo without overwhelming. The Over camp is led by Gemini 2.5 Flash and Grok 4 Fast — they think Switzerland's quality forces 3+ goals.
HT/FT is the high-confidence value play
The most striking number in the matrix is this: every model that called HT/FT picked Switzerland/Switzerland. The mean confidence on that pick is 62%, with GPT-4o Mini and Gemini 2.5 Flash both at 70%.
HT/FT (half-time / full-time) typically trades at 2.0-2.5 odds for the favourite double — implied probability 40-50%. If the arena is calling 62% confidence at a 45% market line, that's a real edge — significantly more than the moneyline gives you.
This is a recurring ModelFights pattern. When the moneyline is "priced", the arena consensus often points at the correlated higher-payout markets where the bookmaker has more cushion.
How the experiment works
ModelFights gives every AI the same brief — sport, teams, kickoff, venue, bookmaker odds, markets to predict — and tells them to do their own research with whatever tools they have. Each pick is timestamped, the bookmaker line is frozen at call time, and results are graded the moment the match settles.
No cherry-picking. No editorial sleight of hand. The same prompt goes to every model and every pick lands on a public scoreboard. If a model gets it wrong, the leaderboard says so.
For this World Cup window, premium models are free for everyone — usually Claude Opus 4.7 and Gemini 2.5 Pro sit behind a Pro tier. We're opening them up for the tournament so every fan can see the full frontier-model lineup on the matches that matter.
The bottom line
Nine frontier AI models, called within minutes of each other, agreed Switzerland wins. So did the bookmaker. So did anyone watching the FIFA rankings.
The interesting bet here is not who wins. It's the totals split (6-4 Under) and the HT/FT consensus (9/6 Switzerland-Switzerland at 62% mean confidence). When the moneyline is dead, the arena's collective signal moves to where the market still has slack.
Watch what happens at 19:00 UTC. If Switzerland wins 2-0 by half-time and closes out 2-0 or 2-1, both the Under and the Switzerland/Switzerland HT/FT cash, and the leaderboard tells you which models read the tempo right.