Premier League Week 27 AI Model Prediction Audit
Mistral Small 3.2 24B led Premier League predictions with 2.11 points per match, followed by Devstral 2 (1.89) and Gemma 3 27B (1.56). Models achieved 41.52% correct tendency overall, though Aston Villa's 1-1 draw with Leeds caught 94.7% consensus wrong.
Mistral Small 3.2 24B led Premier League predictions with 2.11 points per match, followed by Devstral 2 (1.89) and Gemma 3 27B (1.56). Models achieved 41.52% correct tendency overall, though Aston Villa's 1-1 draw with Leeds caught 94.7% consensus wrong.
Premier League Regular Season - 27 featured 9 matches including high-profile fixtures like Tottenham vs Arsenal and Manchester City vs Newcastle. AI prediction accuracy remains critical as models compete on tendency and exact score precision across varied match dynamics. Statistical performance highlights follow.
Top 10 Models
| # | Model | Matches | Total Points | Avg Pts/Match | Tendency % | Exact % |
|---|---|---|---|---|---|---|
| 1 | Mistral Small 3.2 24B (OpenRouter) | 9 | 19 | 2.11 | 44.4% | 11.1% |
| 2 | Devstral 2 (OpenRouter) | 9 | 17 | 1.89 | 55.6% | 11.1% |
| 3 | Gemma 3 27B (OpenRouter) | 9 | 14 | 1.56 | 44.4% | 11.1% |
| 4 | Trinity Large Preview (OpenRouter) | 9 | 13 | 1.44 | 55.6% | 0.0% |
| 5 | DeepSeek R1-0528 (OpenRouter) | 9 | 13 | 1.44 | 44.4% | 0.0% |
| 6 | Qwen3 30B A3B (OpenRouter) | 9 | 13 | 1.44 | 44.4% | 11.1% |
| 7 | Llama 3.3 70B Instruct (OpenRouter) | 9 | 12 | 1.33 | 33.3% | 11.1% |
| 8 | Phi-4 (OpenRouter) | 9 | 11 | 1.22 | 33.3% | 11.1% |
| 9 | GLM-4.7 (OpenRouter) | 9 | 11 | 1.22 | 55.6% | 0.0% |
| 10 | DeepSeek V3.2 (OpenRouter) | 9 | 11 | 1.22 | 33.3% | 11.1% |
Match-by-Match Audit
- Tottenham vs Arsenal (1-4): 84.2% correct tendency, 0.0% exact hits. Consensus: A (84.2%) correct.
- Nottingham Forest vs Liverpool (0-1): 63.2% correct tendency, 0.0% exact hits. Consensus: A (63.2%) correct.
- Sunderland vs Fulham (1-3): 0.0% correct tendency, 0.0% exact hits. Consensus: D (68.4%) incorrect.
- Crystal Palace vs Wolves (1-0): 63.2% correct tendency, 0.0% exact hits. Consensus: H (63.2%) correct.
- Manchester City vs Newcastle (2-1): 84.2% correct tendency, 42.1% exact hits. Consensus: H (84.2%) correct.
- West Ham vs Bournemouth (0-0): 57.9% correct tendency, 0.0% exact hits. Consensus: D (57.9%) correct.
- Chelsea vs Burnley (1-1): 15.8% correct tendency, 0.0% exact hits. Consensus: H (84.2%) incorrect.
- Aston Villa vs Leeds (1-1): 5.3% correct tendency, 0.0% exact hits. Consensus: H (94.7%) incorrect.
- Brentford vs Brighton (0-2): 0.0% correct tendency, 0.0% exact hits. Consensus: H (73.7%) incorrect.
Biggest Consensus Misses
- Aston Villa vs Leeds (1-1): Consensus H (94.7%) incorrect
- Chelsea vs Burnley (1-1): Consensus H (84.2%) incorrect
- Brentford vs Brighton (0-2): Consensus H (73.7%) incorrect
- Sunderland vs Fulham (1-3): Consensus D (68.4%) incorrect
Methodology
kroam.xyz uses a quota-based scoring system that rewards both accuracy and boldness:
Tendency Points (2-6 points): Models earn points for correctly predicting the match outcome (home win, draw, or away win). The points awarded depend on prediction rarityβif most models predicted a home win but the away team won, models who correctly predicted the away win earn more points (up to 6). Common predictions earn fewer points (minimum 2).
Goal Difference Bonus (+1 point): If the model predicts the correct goal difference (e.g., predicted 2-1 and result was 3-2, both +1 difference), they earn a bonus point.
Exact Score Bonus (+3 points): Predicting the exact final score earns 3 additional points.
Maximum: 10 points per prediction (6 tendency + 1 goal diff + 3 exact).
This system ensures that models taking calculated risks on unlikely outcomes are rewarded when correct, while also recognizing precision in exact score predictions. Learn more about our methodology.
Frequently Asked Questions
Q: Which AI model performed best in Premier League Regular Season - 27? A: Mistral Small 3.2 24B achieved the highest average points per match (2.11) across all 9 fixtures.
Q: How accurate were AI predictions for Premier League this round? A: Models achieved 41.52% correct tendency and 4.68% exact score hit rate across 171 total predictions.
Q: What was the biggest upset in Premier League Regular Season - 27? A: Aston Villa vs Leeds (1-1) was the biggest consensus miss, with 94.7% of models incorrectly predicting a home win.
Q: How does kroam.xyz score AI football predictions? A: The quota-based system awards 2-6 points for correct tendency, +1 for goal difference, and +3 for exact scores, with a maximum of 10 points per prediction.
Generation cost: $0.0020
Tokens: 4,832 input + 1,697 output
Frequently Asked Questions
What is this article about?
Which AI model performed best in Premier League Regular Season - 27?**?
Q: Which AI model performed best in Premier League Regular Season - 27?
Q: How accurate were AI predictions for Premier League this round?
You might also like
Premier League Round 26 AI Model Predictions Audit
GLM 4.7 Flash led Premier League predictions this week with 4.67 points per match, followed by DeepSeek R1 0528 (4.63) and Kimi K2 Instruct (3.88). Models achieved 46.15% correct tendency overall, though Chelsea's 2-2 draw with Leeds caught most models off guard.
Feb 16, 2026
UEFA Europa League Round of 32 AI Model Performance Audit
GLM-5 (OpenRouter) led UEFA Europa League predictions this week with 3.25 points per match, followed by Llama 4 Scout (OpenRouter) at 2.88 and Mistral Small 3.2 24B (OpenRouter) at 2.25. Models achieved 52.63% correct tendency overall, though Ludogorets vs Ferencvarosi TC (2-1) caught most models off guard.
Feb 23, 2026
UEFA Conference League Round of 32 AI Prediction Audit
Trinity Large Preview led with 3.13 points per match, followed by Phi-4 (2.38) and Kimi K2.5 (2.13). Models achieved 33.19% correct tendency overall, with FC Noah's 1-0 win over AZ Alkmaar being the biggest surprise.