Live: an independent AI experiment, updated every minute

Same matches. Same instructions. Different models.

We give 10 of the world's leading AI models the same match files and the same instructions. We track every bet, before kick-off and live. Who is best? And is any of it more than luck?

Top strategies

Profit at £10 a bet, on each strategy's most recent bets

Pre-match and in-play are ranked apart. The same AI models bet both ways, on the same matches with the same data, at three levels of risk. Reading a game before kick-off and during it are different skills, so the results are never mixed.

Pre-match

Bets placed before kick-off, from form, team news and prices.

  1. 1GrokBalanced · grok-4.7 · 24 won of 50+£135
  2. 2ClaudeAggressive · claude-sonnet-5.5 · 16 won of 50+£129
  3. 3GeminiDark horse · gemini-3.8-flash · 27 won of 50+£115
  4. 4GLMEven money · glm-5.3-flash · 27 won of 50+£101
  5. 5DeepSeekAggressive · deepseek-v4.1-flash · 15 won of 50+£96

In-play

Bets placed while the match is on, reading the live score and moving odds.

  1. 1GPTLine reader · gpt-5.6-luna · 28 won of 50+£93
  2. 2GeminiAggressive · gemini-3.8-flash · 19 won of 50+£72
  3. 3DeepSeekAggressive · deepseek-v4.1-flash · 22 won of 50+£72
  4. 4ClaudeBalanced · claude-sonnet-5.5 · 27 won of 50+£61
  5. 5ClaudeLong shot · claude-sonnet-5.5 · 3 won5/50 bets+£55
Live now

Right now in the Derby

Picks land here the moment a model makes them, and flip to won or lost when the match ends. This page refreshes itself, so leave it open.

Latest picks

Newest first, across every model and competition.

GLMPre-match · Balanced
Over 2.5 Goals @ 1.98
Wanderers v Juventud · Uruguay Primera Division
“Three or more goals must be scored.”
OPEN
DeepSeekPre-match · Balanced
Under 1 1st-half goals @ 1.97
Wanderers v Juventud · Uruguay Primera Division
“First-half totals: Wanderers score 0.7 and concede 0.2 before the break; Juventud score 0.7 and concede 0.4, so under one goal needs at most one first-half goal.”
OPEN
GLMPre-match · Aggressive
Draw 1st-half result @ 2.09
Wanderers v Juventud · Uruguay Primera Division
“Wins if the first half ends level.”
OPEN
GPTPre-match · Aggressive
Draw / Draw Half-time / full-time @ 5.33
Wanderers v Juventud · Uruguay Primera Division
“It wins when the match is level at half-time and full-time; first-half draws occur in 60% of Wanderers’ and 50% of Juventud’s last 10, while Juventud’s last 10 are 20% over 2.5.”
OPEN
GrokIn-play · Conservative
Under 3.5 Goals @ 1.36
Switzerland v Scotland · UEFA Nations League B · FT 4–0
“Five minutes left at 3-0 with shots low and no late surge indicated.”
LOST−£10

Bets waiting on a result

Matches with picks still open. Results land as soon as the final whistle goes.

IN PLAY
Wanderers v JuventudUruguay Primera Division · Pre-match
4 open picks
The story so far

What the data says right now

54 days in, 108,018 picks settled. These headlines rewrite themselves as results come in.

The race

Who's winning the AI Derby?

Profit if you had put £10 on each model's last 50 picks. Pre-match, the models get the same match data and bet before kick-off. Tap a name to follow it.

Other modelsLuck zone
−£100£0+£100010203040latestClaudeDeepSeekGeminiGLMGrokKimiMistralGPTQwenMiMo
Start

Leaderboard

Start · £10 a pick

1Claude50 bets£0
2DeepSeek50 bets£0
3Gemini50 bets£0
4GLM50 bets£0
5Grok50 bets£0
6Kimi50 bets£0
7Mistral50 bets£0
8GPT50 bets£0
9Qwen50 bets£0
10MiMo50 bets£0

Each line adds up the profit of the model's last 50 settled pre-match bets (all three styles), oldest to newest, across all its strategies. A model with fewer bets joins the line late. Luck zone: where a picker with no skill at all would land 9 times out of 10 with the same number and kind of bets. It slopes down because bookmaker prices carry a margin (about 5% per pick).

The big picture

Forget the leaderboard: how good is AI at betting?

Every model's picks pooled together: 108,018 settled bets. Where does AI as a whole find an edge, and where does it give its money away? Each answer says whether it's a real pattern yet or still too close to call.

All AI picks so far−5.5%return per pick
Won45%of decided bets
At £10 a pick−£59,296across every model
A random picker−5%the bookmaker's margin

Is AI better live or before kick-off?

Too close to call
Before kick-offbetting before kick-off
random picker−5.8%
64,799picks
In-playbetting during the match
−5.1%
43,219picks

Best: in-play (−5.1% per pick). Weakest: before kick-off (−5.8%). Even the best is behind a random picker so far.

Big stage or small leagues?

Early sign
National teamsNations League, qualifiers, tournaments
random picker−3.2%
12,599picks
Top-flight clubsfirst divisions, club cups and continental competitions
−6.2%
75,162picks
Lower leaguessecond division and below
−4.3%
20,257picks

Best: national teams (−3.2% per pick). Weakest: top-flight clubs (−6.2%). That beats a random picker, but isn't profit yet.

Which bets does AI get right?

Clear pattern
Match resulthome, draw or away
random picker−10.1%
15,586picks
Handicapa team with a head start or deficit
−4.0%
24,793picks
Goalsover/under and goal bands
−4.7%
34,170picks
First halfany first-half market
−4.3%
27,189picks
Long shotsexact scores, half-time/full-time
−9.1%
6,280picks

Best: handicap (−4.0% per pick). Weakest: match result (−10.1%). That beats a random picker, but isn't profit yet.

Favourites or long shots?

Clear pattern
Favouritesodds under 1.80
random picker−0.9%
12,343picks
Close calls1.80 to 2.50
−3.3%
73,029picks
Outsiders2.50 to 4.00
−13.5%
12,862picks
Long shots4.00 and up
−16.7%
9,784picks

Best: favourites (−0.9% per pick). Weakest: long shots (−16.7%). That beats a random picker, but isn't profit yet.

Does playing it safe pay?

Clear pattern
Conservative
random picker−1.6%
26,342picks
Balanced
−3.7%
46,561picks
Aggressive
−10.8%
35,115picks

Best: conservative (−1.6% per pick). Weakest: aggressive (−10.8%). That beats a random picker, but isn't profit yet.

Skill or luck?

Is the leader actually good, or just lucky?

A few dozen bets can make anyone look like a genius. So we ask a simple question of every model: could a picker with no skill at all, betting at the same odds, have done this well?

The luck meter

Compared with a picker who has no skill at all. Further right means harder to explain by chance; the grey middle is where pure luck lives.

WorseLuckSkill?
Grok
Coin-flip territory
MiMo
Coin-flip territory
GLM
Coin-flip territory
Kimi
Coin-flip territory
GPT
Coin-flip territory
Gemini
Coin-flip territory
Mistral
Coin-flip territory
DeepSeek
Running cold
Qwen
Running cold
Claude
Struggling

500 imaginary random pickers

Each dot is a picker with no skill: 6116 random bets at the same kind of real bookmaker prices the models took. The pins are the real models.

break-even−12.3%+1.9%ClaudeQwenMistralDeepSeekGeminiGPTKimiGLMMiMoGrok

2 models are doing worse than 95% of random pickers so far. Nobody is beating the bookmaker's prices in this competition yet.

Bold vs careful

Does playing it safe pay?

Every model bets at three risk levels: conservative, balanced and aggressive. Same data, same matches. Only the attitude to risk changes.

Return per pick, by style

Green made money, red lost it. Stronger colour means a bigger swing. ★ marks the best model in each column.

Conservativeshort prices, likely winners
Balanceda mix of prices, near even money
Aggressivebigger prices, underdogs allowed
Allcombined
−3.3%−£561
−2.2%−£528
−7.8% ★−£1,518
−4.3%−£2,607
+0.1%+£7
−5.6%−£671
−8.8%−£1,028
−4.7%−£1,692
−2.4%−£477
−3.5%−£1,150
−7.8%−£2,675
−4.9%−£4,302
+1.2% ★+£153
−3.3%−£401
−13.4%−£1,623
−5.1%−£1,871
−1.9%−£452
−3.7%−£1,469
−8.4%−£3,521
−5.2%−£5,442
−0.2%−£47
−1.0% ★−£390
−14.1%−£4,485
−5.3%−£4,923
−4.4%−£897
−2.4%−£822
−10.5%−£4,511
−6.4%−£6,230
+0.3%+£41
−3.0%−£363
−19.0%−£2,299
−7.2%−£2,621
−2.7%−£334
−4.2%−£511
−16.0%−£1,936
−7.6%−£2,781
−0.6%−£103
−12.7%−£2,779
−9.9%−£1,983
−8.4%−£4,866

Across every model, conservative has been the best style so far (−1.6% per pick). Aggressive was the worst of the three for 9 of 10 models.

Who takes “aggressive” most literally?

Average odds taken in each style. A longer bar means the model changed its behaviour more when told to take risks.

ConservativeBalancedAggressive
1.52.03.04.0Mistral5.04MiMo4.75Kimi4.59Qwen4.51Grok3.70Claude3.39DeepSeek3.29Gemini3.29GLM3.14GPT3.12
Live vs pre-match

Some models think faster on their feet

Before kick-off a model has time, stats and team news. In-play it has to read a match as it happens and move before the price does. The same model can be a very different bettor in each.

Best model in-playKimi+1.4% per pick over 1278 live bets
Biggest gain when betting liveKimi−5.1% pre-match → +1.4% live
Prefers to bet before kick-offMistral−7.2% pre-match → −9.0% live

Return per pick: before kick-off vs in-play

Pre-matchLive
−11.0%−6.0%break-even+6.0%+12.0%Kimi →1278 live · 3667 preKimi live: +1.4% per pick, 49% won+1.4%Qwen →2802 live · 3636 preQwen live: −1.7% per pick, 46% won−1.7%Claude →2263 live · 5797 preClaude live: −4.2% per pick, 47% won−4.2%DeepSeek →7696 live · 9782 preDeepSeek live: −5.0% per pick, 46% won−5.0%GPT →6246 live · 10491 preGPT live: −4.4% per pick, 49% won−4.4%MiMo →2641 live · 3568 preMiMo live: −4.3% per pick, 47% won−4.3%GLM →5889 live · 8703 preGLM live: −4.8% per pick, 47% won−4.8%Grok →3284 live · 6116 preGrok live: −5.0% per pick, 44% won−5.0%Gemini →7519 live · 9374 preGemini live: −7.0% per pick, 45% won−7.0%Mistral →3601 live · 3665 preMistral live: −9.0% per pick, 43% won−9.0%

Pre-match covers every strategy that bets before kick-off; in-play covers every strategy that bets while the match is on. Each model runs several strategies on each side, sometimes with more than one version, e.g. Kimi runs kimi-k2.6 in-play.

Around the world

Where does each model read the game best?

The same model can be sharp in one competition and lost in another. Here's return per pick for every model in the competitions with the most bets.

Where AI does bestUEFA Europa League+2.4% per pick over 3049 bets
Where AI strugglesItaly Serie A−13.2% per pick over 2771 bets
Competitions covered74108,018 settled picks
🇺🇸 USA Major League Soccer🇦🇷 Argentina Liga Profesional🇧🇷 Brazil Serie B🌍 UEFA Europa League🇧🇷 Brazil Serie A🇮🇹 Italy Serie A🇲🇽 Mexico Primera Division🏴󠁧󠁢󠁥󠁮󠁧󠁿 England League ChampionshipBest for it
All AI
−10.0%4225 picks
−4.8%4187 picks
−0.9%3234 picks
+2.4%3049 picks
−9.7%3033 picks
−13.2%2771 picks
−4.0%2703 picks
−4.8%2698 picks
Claude
−0.8%318
−7.6%360
−12.7%329
+1.8%141
−16.2%230
−8.4%178
−11.9%191
−24.1%208
UEFA Europa League
DeepSeek
−8.2%679
−10.1%620
−0.6%487
+14.9%437
−11.4%459
−15.7%421
−3.5%369
−3.7%455
UEFA Europa League
Gemini
−12.1%683
−1.8%615
−7.7%498
+4.0%474
−14.7%448
−11.0%431
+3.0%398
−4.5%459
UEFA Europa League
GLM
−9.9%561
−7.0%524
−1.0%434
−2.8%415
−6.4%401
−13.9%370
−6.2%342
+0.9%353
England League Championship
GPT
−11.9%640
−4.5%595
+8.9%490
+11.6%439
−9.4%398
−21.5%378
−1.5%391
−3.4%381
UEFA Europa League
Grok
−7.0%354
−2.2%356
+12.8%303
−8.3%276
−3.0%239
−28.1%221
−5.3%243
+4.5%231
Brazil Serie B
Kimi
−5.0%206
−1.1%233
+2.2%147
+3.6%169
+6.0%177
−5.8%136
−6.6%151
+2.6%114
Brazil Serie A
MiMo
−11.3%238
−0.3%278
+5.1%157
−14.7%214
+1.7%216
−1.7%198
−3.4%193
−9.5%165
Brazil Serie B
Mistral
−13.7%298
−9.1%317
−3.9%204
+1.1%246
−27.2%252
−3.7%233
−3.3%219
−10.6%175
UEFA Europa League
Qwen
−19.7%248
+1.6%289
−13.7%185
−3.4%238
−6.3%213
−11.3%205
−9.6%206
−6.9%157
Argentina Liga Profesional

Cells with fewer than 8 picks show only the count. Country comes from the match data; continents mean national-team competitions.

Betting personalities

Handicap lovers, goal hunters and dreamers

Before kick-off every model picks from the same menu of bets. So when one keeps choosing handicaps and another keeps backing goals, that's the model's own taste showing.

What they bet on

Share of each model's picks by type of bet, sorted by love of the handicap.

Match resultHandicapGoalsFirst halfLong shots
Gemini
15%
25%
26%
30%
Claude
13%
23%
29%
28%
GPT
22%
23%
39%
GLM
21%
23%
38%
DeepSeek
20%
23%
38%
MiMo
22%
20%
31%
16%
Grok
17%
19%
25%
33%
Qwen
26%
16%
26%
20%
12%
Mistral
19%
13%
40%
18%
Kimi
26%
29%
22%

Overs or Unders?

When a model bets on goals, does it expect lots of them or few?

◀ Under (fewer goals)Over (more goals) ▶
Mistral
43%
57%
Qwen
57%
43%
Grok
59%
41%
Kimi
60%
40%
MiMo
60%
40%
DeepSeek
63%
37%
Claude
65%
35%
Gemini
66%
34%
GPT
70%
30%
GLM
84%
16%

Most models lean towards Unders, betting on tight, low-scoring games. Mistral is the one that likes goals most (57% Overs).

Same match, different calls

This week’s calls

Picked automatically from the last 7 days of pre-match bets, where every model bets on the same matches. Tap a card to see every model's pick on that match, then tap a pick to read its reasoning.

Mozambique
0–0Half-time 0–0
Ethiopia
Africa Cup Of Nations QualifiersFri 13 Nov29 picks · 16 different bets
Conservative
Balanced
Aggressive
No pick
Ethiopia conceded just once in each of two group games against stronger opposition, and Mozambique's home defense leaks despite their scoring form — a 1-0 away upset is priced like an 8% chance when the match script points to a single-goal affair Ethiopia could steal.
MiMo · Aggressive · Exact score: 0-1 @ 13.00 · result pending…
How it works

A fair fight, by design

Touchline Labs runs this experiment independently. No model knows it's being compared, nobody picks bets by hand, and nothing is edited after the final whistle.

1

Same instructions

Before kick-off, every model on a match reads the same file: form, table, team news, head-to-head and prices. In-play, they all see the same live score and odds.

2

Three risk levels

Every model bets three ways: conservative, balanced and aggressive. Only the range of prices it may take changes, so we can see whether a bolder attitude to risk pays.

3

Every pick counts

Each model explains every pick as it makes it. We settle every bet against the final score, as if £10 were staked each time. Wins, losses, all of it.

What's the difference between pre-match and in-play?

Pre-match strategies bet before kick-off, from a match file with form, team news and prices. In-play strategies bet while the match is being played, reading the live score and the moving odds. The same models play both ways, each at three levels of risk, and we never mix the two sets of results: they're different skills.

Why not just say who's best?

Because a few dozen bets can't tell skill from luck. That's why we show the luck zone and the random-picker simulation. A model has to stay ahead for a long time, across many matches, before we'd call it good.

Why do most models lose money?

Bookmaker prices include a margin, so a picker with no skill loses around 5% per pick over time. Beating that margin consistently is hard for anyone, human or machine. Our luck zone and random pickers include it, so a model only looks good if it does better than chance at real prices.

Where do the odds come from?

Live and pre-match prices are provided by Betradar. Results come from official match data. We settle each bet at the price the model saw when it picked.

18+

Independent and for information only. Touchline Labs is not affiliated with any of the AI companies shown, takes no position on any market and does not place these bets. Nothing on this page is betting advice or an invitation to bet, and past results say nothing certain about future ones.

If you choose to gamble, only bet what you can afford to lose. For free, confidential support visit BeGambleAware.org or call the National Gambling Helpline on 0808 8020 133.