— Live · updated every 6 hours
AI prediction accuracy report
10 AI models, 1,348 games, 12,209 graded picks (1 Oct – 31 Oct 2026).
Every pick is made before the game and graded after it. Wins and losses both count.
Key findings
- Meta Muse Spark 1.3 picked the right winner most often: 55.8% of 77 picks. MiMo V2.6 Pro was last at 51.8%, so first to last was only 4.0 points.
- When every model agreed (755 of 1,348 games), the pick was right 60.7% of the time, against 54.6% for the majority pick across all games.
- Confidence means something: picks made with 70% or more confidence were right 73.6%, picks under 50% only 42.6%.
- Every model backed the home team more often than home teams won: 48.1% to 68.8% of picks, against 47.3% of games won at home.
- In football (soccer), where draws make it a three-way call, the majority of the models was right 53% of 782 matches, against 43.4% for always backing the home team and 33.3% for a random pick.
Accuracy by AI model
Share of picks where the model called the winner (or the draw) correctly, all sports.
| # | Model | Right winner | Picks | Exact score | Backed home team |
|---|---|---|---|---|---|
| 1 | Meta Muse Spark 1.3 |
55.8%
|
77 | 11.7% | 48.1% |
| 2 | Kimi K3 |
54.4%
|
1,348 | 7.9% | 54.2% |
| 3 | Claude Sonnet 5.5 |
54.3%
|
1,348 | 6.8% | 68.8% |
| 4 | Gemini 3.8 Flash |
54.2%
|
1,348 | 7.8% | 52.1% |
| 5 | GLM 5.3 |
53.9%
|
1,348 | 6.8% | 59.6% |
| 6 | GPT-6 Luna |
53.9%
|
1,348 | 7.5% | 53.3% |
| 7 | Grok 4.7 |
53.8%
|
1,348 | 8.2% | 59.1% |
| 8 | Qwen 3.8 Max |
52.5%
|
1,348 | 7.3% | 53.6% |
| 9 | DeepSeek V4 Pro |
52.4%
|
1,348 | 6.8% | 55.1% |
| 10 | MiMo V2.6 Pro |
51.8%
|
1,348 | 6.8% | 56.2% |
When the models agree
- All models agreed (755 games)
- 60.7%
- Majority pick, all games
- 54.6%
- Best single model (Kimi K3)
- 54.4%
- Home team won
- 47.3%
Does AI confidence mean anything?
- Stated confidence under 50% (2,654 picks)
- 42.6% right
- Stated confidence 50–59% (6,450 picks)
- 54.3% right
- Stated confidence 60–69% (2,760 picks)
- 59.6% right
- Stated confidence 70% and up (345 picks)
- 73.6% right
Accuracy by sport
Sports with fewer than 30 graded games are left out of this table but count in the totals.
| Sport | Games | Majority right | All agreed, right | Home won | Best model |
|---|---|---|---|---|---|
| Football (soccer) | 782 | 53% | 62.4% (428) | 43.4% | Gemini 3.8 Flash · 54.6% |
| Ice hockey | 285 | 60% | 62.9% (167) | 57.2% | Kimi K3 · 61.4% |
| NBA | 146 | 55.5% | 58.3% (84) | 53.4% | GLM 5.3 · 58.2% |
| Baseball | 75 | 46.7% | 48.5% (33) | 45.3% | Qwen 3.8 Max · 54.7% |
| NFL | 60 | 56.7% | 48.8% (43) | 40% | GPT-6 Luna · 56.7% |
Use the data
Free to republish under CC BY 4.0. Please credit TuringStats with a link to this page.
Cite as: TuringStats AI Prediction Accuracy Report (last 30 days to 31 October 2026), turingstats.com/ai-accuracy-report
Download every graded pick (CSV)Embed the live leaderboard
Paste this into any article or blog. It updates on its own.
Method
- All 10 models get the same match data for every game (form, venue, head-to-head and, where available, team news), never the betting odds. When the prompt changes, it changes for every model at once.
- Each model returns a winner (or a draw where the sport allows it), a score and a confidence from 0 to 100. Only its latest pick before the game counts.
- A pick is right when the result matches. "Majority" is the result most models picked (ties are left out); "all agreed" means every model picked the same result.
- Results come from official match data. The line-up has been fixed since 7 September 2026; earlier picks are excluded.
FAQ
Which AI is best at predicting sports?
Meta Muse Spark 1.3 picked the right winner most often: 55.8% of 77 picks. MiMo V2.6 Pro was last at 51.8%, so first to last was only 4.0 points.
Are AI sports predictions accurate?
Better than chance, far from certain. Across 1,348 games the majority pick of the models was right 54.6% of the time, and 60.7% when every model agreed.
How is the accuracy measured?
Every model predicts the winner (or a draw), the score and its confidence before kick-off. After the final whistle each pick is graded right or wrong. Only the latest pick before the game counts, and every graded pick is in the CSV download.
Can I use these numbers?
Yes. The report and data are free to use under CC BY 4.0: credit "TuringStats" with a link to this page.