Which AI predicts sports best?

quasiperfectus1 pts0 comments

AI Sports Prediction Leaderboard — Which AI Predicts Best? | Predicted Sports

Sign in<br>Sign up

The AI Sports Prediction Index

Which AI predicts sports best?

GPT-5.5, GLM 5.2, Claude Opus 4.8, Gemini 3.5 Flash, Grok 4.3, DeepSeek V4 Pro, and Claude Sonnet 5 — head to head on real games. Each gets the identical line-blind data packet (no betting line, no web search), locks its call ~3 hours before first pitch, and is graded in public. No edits, no do-overs.

The honest finding so far: nobody beats the closing line consistently — not even the frontier models. The race worth watching is who gets closest.

See the standings<br>How it's scored<br>Can AI beat Vegas? →<br>Our own models' record →

🥇

GLM 5.2

43–30

59% · Brier 0.236

🥈

DeepSeek V4 Pro

55–38

59% · Brier 0.242

🥉

GPT-5.5

63–42

60% · Brier 0.243

MLB standings · ranked by Brier (lower = better)

105 graded games

Model<br>Record<br>AccAccuracy<br>Brier<br>ROI<br>Games

GLM 5.2👑

43–30<br>59%<br>0.236

+6.3%

73

DeepSeek V4 Pro

55–38<br>59%<br>0.242

+7.8%

93

GPT-5.5

63–42<br>60%<br>0.243

+8.8%

105

Claude Opus 4.8

57–48<br>54%<br>0.244

-2.0%

105

Claude Sonnet 5

52–33<br>61%<br>0.246

+12.6%

85

Gemini 3.5 Flash

57–48<br>54%<br>0.246

-2.7%

105

Grok 4.3

56–49<br>53%<br>0.250

-4.7%

105

Pick the favoritebaseline

62–41<br>60%<br>0.398

+7.0%

103

Pick the home teambaseline

48–57<br>46%<br>0.543

-14.9%

105

Brier = mean squared error of the win probability (0 = perfect, 0.25 = a coin flip). Accuracy = how often the model's side won. ROI = return on a 1-unit bet on each pick at the closing market price (so beating the vig means clearing ~0%). The baselines are the bar: an AI that can't out-forecast "pick the home team" isn't forecasting. Early samples are small — read with care.

Get our sports prediction daily newsletter

Featuring our models and analysis — the AI board, our own picks, and where they disagree. Free, in your inbox every morning.

Leave this field empty

Run totals · over/under, ranked by accuracy

Model<br>O/U acc<br>Avg miss<br>ROI<br>Totals graded

Grok 4.3<br>64%<br>±3.9

+23.0%

102

GLM 5.2<br>61%<br>±4.2

+21.1%

72

DeepSeek V4 Pro<br>60%<br>±4.0

+14.6%

90

GPT-5.5<br>58%<br>±4.0

+11.5%

102

Claude Opus 4.8<br>57%<br>±4.0

+9.1%

102

Gemini 3.5 Flash<br>55%<br>±4.0

+4.6%

102

Claude Sonnet 5<br>54%<br>±4.1

+6.3%

83

Market (closing line)baseline<br>57%

+6.6%

102

Each model projects the game's total runs (line-blind); we grade its over/under call against the closing market line (a no-vig multi-book consensus via The Odds API where captured; Kalshi before that). O/U acc = how often that call was right. Avg miss = mean runs off the actual total. ROI = return on a 1-unit bet at the closing price — so break-even is the market's price (here the favored side runs ~59¢), not 50%: a model can top 50% and still lose to the vig. The Market (closing line) row is the bar to clear. Early samples are small — read with care.

🔒 See each model's projected total on every game — not just the scoreboard.

Start 7-day free trial<br>Cancel anytime · then $14/mo

First inning · run in the 1st (YRFI), ranked by Brier

Model<br>YRFI acc<br>Brier<br>Graded

Gemini 3.5 Flash<br>54%<br>0.240<br>69

Claude Sonnet 5<br>54%<br>0.244<br>68

DeepSeek V4 Pro<br>59%<br>0.244<br>61

GPT-5.5<br>58%<br>0.244<br>69

Grok 4.3<br>61%<br>0.245<br>69

GLM 5.2<br>48%<br>0.249<br>69

Claude Opus 4.8<br>49%<br>0.250<br>45

Always “no run”baseline<br>44%<br>0.557<br>70

Each model calls the probability of a run in the first inning (either team), line-blind. First innings are close to a coin weighted toward "no run", so the Always "no run" baseline is the bar — an AI only shows skill by beating it on Brier. It's also the one market where our own first-inning model claims a real (modest) edge, so this table is the fairest fight on the board. Early samples are small — read with care.

UFC — best fight forecasters

Same idea, in the cage: every model gets the identical line-blind fight packet — no odds, no search — calls the winner and the method , and is graded on both. Follow the picks on every card.

No graded fights yet. Each model's locked picks grade as fights resolve — the board fills from the next card on.

How it works

Line-blind. No model ever sees the betting line. Each produces its own win probabilities and run projections from the data alone, so the board reads forecasting skill — not an echo of the market.

One packet, no search. Every model gets the same point-in-time data (ratings, Statcast, bullpens, park/umpire, situational splits) ~3h before first pitch. No web search, so it's the models we're measuring.

Graded in public, no do-overs. Each model makes one call per game and we live with it — against the outcome (accuracy, Brier) and against the closing price (ROI). Our own house models get the same treatment on the accuracy page.

Where the Index goes next

✓Live now: nightly MLB board, UFC fight board, run totals graded against the closing line, ROI at closing prices.

✓World Cup: our panel is calling every knockout match — follow the bracket.

◌The daily board update — who's hot, who's slipping, where the models disagree,...

model line brier closing claude graded

Related Articles