AI predictions for Premier League matches
AI predictions for Premier League matches are published here before kick-off. Each prediction shows the bet of the largest group of AI models and the bookmaker odds at the time of publication. After the final whistle we settle every bet, win or lose.
The full match breakdown covers team form, league position, head-to-head record, injuries, suspensions and the betting line.
2026/27 season · updated 12 October
How our Premier League predictions did
Our pick since 10 October: 5 of 10 won · ROI −3.5% · result −0.35 units
4:1Coventry City–Newcastle UnitedOur pick:2prediction lost11 of 16 AI models picked this bet1 3.31X 3.592 2.130:3Liverpool–Manchester CityOur pick:2prediction won8 of 15 on this bet · 1 more in the same direction1 3.14X 3.722 2.183:1Crystal Palace–Nottingham ForestOur pick:BTTSprediction won11 of 15 AI models picked this bet1 2.79X 3.342 2.552:2Hull City–EvertonOur pick:U 2.5prediction lost14 of 16 AI models picked this bet1 3.65X 3.322 2.111:1Manchester United–Tottenham HotspurOur pick:1prediction lost8 of 16 on this bet · 2 more in the same direction1 1.74X 4.152 4.402:2Aston Villa–BrentfordOur pick:BTTSprediction won8 of 16 AI models picked this bet1 2.77X 3.512 2.472:1Ipswich Town–FulhamOur pick:U 2.5prediction lost9 of 15 AI models picked this bet1 2.74X 3.472 2.515:1Chelsea–BournemouthOur pick:BTTSprediction won11 of 16 AI models picked this bet1 1.83X 3.922 4.110:2Sunderland–Brighton & Hove AlbionOur pick:2prediction won11 of 15 on this bet · 1 more in the same direction1 3.08X 3.432 2.312:1Arsenal–Leeds UnitedOur pick:H1 (−1.5)prediction lost13 of 16 on this bet · 1 more in the same direction1 1.42X 4.562 7.92All Premier League results with verified predictions
Premier League in the 2026/27 season
After 60 matches, teams are scoring 3.0 goals per game on average. Over 2.5 goals landed in 57% of games, and both teams scored in 55%.
| Metric | Season 2026/27 | Season 2025/26 |
|---|---|---|
| Goals per match | 2.95 | 2.75 |
| Home wins | 38% | 43% |
| Draws | 32% | 27% |
| Away wins | 30% | 30% |
| Both teams to score | 55% | 56% |
| Over 1.5 goals | 77% | 79% |
| Over 2.5 goals | 57% | 55% |
| Over 3.5 goals | 37% | 28% |
This season — from 60 played matches, last season — from all 380 matches.
Total — goals by both teams in a match; shares are of all matches played.
Who is in form
| Best form | Manchester City | · 15 of 15 points |
|---|---|---|
| Most prolific attack | Brighton & Hove Albion | 18 goals in 6 matches · 3.0 per match |
| Tightest defence | Manchester City | 5 conceded in 6 matches · 0.8 per match |
Form — up to the last five matches, most recent on the right: W — win, D — draw, L — loss; compared per match.
Premier League table
| # | Team | W | D | L | P | Goals | Pts |
|---|---|---|---|---|---|---|---|
| 1 | Manchester City | 6 | 0 | 0 | 6 | 16:5 | 18 |
| 2 | Arsenal | 5 | 0 | 1 | 6 | 10:5 | 15 |
| 3 | Brighton & Hove Albion | 4 | 1 | 1 | 6 | 18:5 | 13 |
| 4 | Brentford | 2 | 4 | 0 | 6 | 12:6 | 10 |
| 5 | Everton | 2 | 4 | 0 | 6 | 8:5 | 10 |
| 6 | Chelsea | 3 | 1 | 2 | 6 | 15:13 | 10 |
| 7 | Leeds United | 2 | 3 | 1 | 6 | 8:5 | 9 |
| 8 | Hull City | 2 | 3 | 1 | 6 | 8:6 | 9 |
| 9 | Liverpool | 2 | 3 | 1 | 6 | 7:7 | 9 |
| 10 | Ipswich Town | 3 | 0 | 3 | 6 | 9:12 | 9 |
| 11 | Newcastle United | 2 | 2 | 2 | 6 | 10:13 | 8 |
| 12 | Crystal Palace | 2 | 1 | 3 | 6 | 9:12 | 7 |
| 13 | Manchester United | 1 | 3 | 2 | 6 | 9:9 | 6 |
| 14 | Coventry City | 2 | 0 | 4 | 6 | 5:11 | 6 |
| 15 | Nottingham Forest | 1 | 2 | 3 | 6 | 5:8 | 5 |
| 16 | Aston Villa | 1 | 2 | 3 | 6 | 6:11 | 5 |
| 17 | Sunderland | 1 | 1 | 4 | 6 | 6:12 | 4 |
| 18 | Bournemouth | 0 | 3 | 3 | 6 | 7:13 | 3 |
| 19 | Tottenham Hotspur | 0 | 3 | 3 | 6 | 3:9 | 3 |
| 20 | Fulham | 0 | 2 | 4 | 6 | 6:10 | 2 |
The leader is highlighted.
Which AI predicts Premier League best
Every model sees the same matches and names one bet each. The ranking is by profit.
| # | AI model | Won | Avg odds | ROI | Profit |
|---|---|---|---|---|---|
| 01 | Xiaomi MiMo | 6 of 9 | 2.10 | +36.4% | +3.28 |
| 02 | ChatGPT | 6 of 10 | 1.91 | +8.5% | +0.85 |
| 03 | Mistral | 6 of 10 | 1.78 | +2.8% | +0.28 |
| 04 |  Dots3-Note | 5 of 9 | 1.79 | +2.4% | +0.22 |
| 05 | Claude | 6 of 10 | 1.79 | −2.1% | −0.21 |
| 06 | DeepSeek | 5 of 10 | 1.99 | −3.5% | −0.35 |
| 07 | DeepSeek Flash | 5 of 10 | 1.94 | −3.5% | −0.35 |
| 08 | Qwen | 5 of 10 | 1.94 | −3.5% | −0.35 |
| 09 |  Sonar | 5 of 10 | 1.74 | −14.9% | −1.49 |
| 10 |  Grok | 4 of 10 | 2.07 | −19.2% | −1.92 |
| 11 |  Gemini | 4 of 10 | 2.10 | −20.9% | −2.09 |
| 12 | Nemotron Ultra | 3 of 8 | 1.94 | −30.5% | −2.44 |
| 13 | MiniMax | 3 of 10 | 2.15 | −33.2% | −3.32 |
| 14 | Kimi | 3 of 10 | 2.16 | −36.6% | −3.66 |
| 15 | Nemotron Super | 3 of 10 | 2.07 | −44.2% | −4.42 |
| 16 |  GLM | 3 of 10 | 2.02 | −44.3% | −4.43 |
| All models combined | 72 of 156 | 1.97 | −13.1% | −20.40 | |
| Model consensusour published prediction | 5 of 10 | 1.94 | −3.5% | −0.35 | |
| Bookmakers’ favouritehome or away at the shorter price | 4 of 10 | 2.13 | −22.6% | −2.26 |
All models combined — every bet of every model added up
Model consensus — one bet per match, from the largest group of models pointing the same way. This is what we publish as our prediction
Bookmakers’ favourite — a bet on the side with the shorter price. A yardstick for comparison, not a prediction of ours
The comparison rows are counted on the same matches; the favourite’s price is the last line before kick-off, and matches without it are left out of that row
“won” excludes voids; ROI is profit against the total staked
On a small sample the order can change noticeably from one matchday to the next