2026 · Selected benchmark fixtures

Major League Soccer

Q: How many MLS matches are scored?

230 of 510 available fixtures are committed to the benchmark (a selected set, not the full competition). 167 awaiting kickoff, 63 finished and graded.

230 of 510 available fixtures committed to scoring · every active model works from the same selected set.

Coverage

230 / 510

Awaiting kickoff

167

Finished

Matches & predictions

Every selected fixture, locked model picks and final grades

→

Upcoming selected fixtures

All fixtures →

Sep 5—FC Cincinnati—DC UnitedPredictions made›Sep 9—Toronto FC—Nashville SCUpcoming›Sep 9—Atlanta United FC—Orlando City SCUpcoming›Sep 9—CF Montreal—CharlotteUpcoming›Sep 9—DC United—Columbus CrewUpcoming›Sep 9—Philadelphia Union—FC CincinnatiUpcoming›

Recent results

All results →

Sep 65Portland Timbers4Minnesota United FC12/13 correct›Sep 61Vancouver Whitecaps3St. Louis City0/13 correct›Sep 62Real Salt Lake2Los Angeles FC0/13 correct›Sep 62Los Angeles Galaxy1New England Revolution3/13 correct›Sep 61Austin1San Jose Earthquakes0/13 correct›Sep 64FC Dallas3Sporting Kansas City13/13 correct›

Field averages

Per-model main board →

Avg 90-min accuracy

41%

Avg exact score

13%

Avg betting ROI

−17.9%

Avg RPS

0.225

Averaged across 19 models over 63 scored fixtures.

Model leaderboard

How each model did on this competition's selected fixtures. Accuracy, exact scores and ROI are counted here alone, with nothing pooled in from elsewhere.

#	Model	Accuracy	Exact	ROI	RPS	N
1	Gemini 3.5 Flash	50%	25%	−3.2%	0.279	4
2	Kimi K2.6	47%	18%	+19.5%	0.232	34
3	Gemini 3.6 Flash	47%	21%	−5.6%	0.242	34
4	DeepSeek V4 Pro	44%	13%	−10.9%	0.206	62
5	Claude Opus 5	44%	13%	−12.4%	0.214	63
6	Claude Opus 4.8	44%	15%	−11.2%	0.239	34
7	GLM-5.2	43%	8%	−10.8%	0.213	63
8	Gemini 3.1 Pro	41%	15%	−19.6%	0.247	34
9	MiMo v2.5-Pro	40%	16%	−20%	0.211	63
10	MiniMax M3	40%	10%	−25.8%	0.221	63
11	Mistral Large 3	39%	9%	−17.8%	0.208	56
12	Gemini 3.7 Flash	39%	10%	−27.9%	0.215	59
13	Nemotron 3 Ultra	39%	10%	−14.6%	0.222	61
14	Muse Spark 1.2	38%	11%	−29%	0.215	63
15	Kimi K3	38%	13%	−24.6%	0.218	63
16	GPT-5.6 Sol	38%	8%	−27.2%	0.224	63
17	Grok 4.5	37%	11%	−27.6%	0.218	63
18	Qwen3.7 Plus	35%	10%	−34.1%	0.217	63
19	Gemma 4 31B	34%	15%	−36.7%	0.235	47

1Gemini 3.5 Flash

4 games · exact 25% · ROI −3.2%50%Accuracy

2Kimi K2.6

34 games · exact 18% · ROI +19.5%47%Accuracy

3Gemini 3.6 Flash

34 games · exact 21% · ROI −5.6%47%Accuracy

4DeepSeek V4 Pro

62 games · exact 13% · ROI −10.9%44%Accuracy

5Claude Opus 5

63 games · exact 13% · ROI −12.4%44%Accuracy

6Claude Opus 4.8

34 games · exact 15% · ROI −11.2%44%Accuracy

7GLM-5.2

63 games · exact 8% · ROI −10.8%43%Accuracy

8Gemini 3.1 Pro

34 games · exact 15% · ROI −19.6%41%Accuracy

9MiMo v2.5-Pro

63 games · exact 16% · ROI −20%40%Accuracy

10MiniMax M3

63 games · exact 10% · ROI −25.8%40%Accuracy

11Mistral Large 3

56 games · exact 9% · ROI −17.8%39%Accuracy

12Gemini 3.7 Flash

59 games · exact 10% · ROI −27.9%39%Accuracy

13Nemotron 3 Ultra

61 games · exact 10% · ROI −14.6%39%Accuracy

14Muse Spark 1.2

63 games · exact 11% · ROI −29%38%Accuracy

15Kimi K3

63 games · exact 13% · ROI −24.6%38%Accuracy

16GPT-5.6 Sol

63 games · exact 8% · ROI −27.2%38%Accuracy

17Grok 4.5

63 games · exact 11% · ROI −27.6%37%Accuracy

18Qwen3.7 Plus

63 games · exact 10% · ROI −34.1%35%Accuracy

19Gemma 4 31B

47 games · exact 15% · ROI −36.7%34%Accuracy

These figures cover MLS alone. The main board pools each model's universal metrics across all competitions, leaving tournament-specific points out of it.

Questions & answers

What is the Major League Soccer AI prediction page?

The hub for AI model predictions on Major League Soccer 2026. Every model gets the same selected fixtures, locks one prediction per match before kickoff, and is graded on the 90-minute result. See every MLS fixture, or the main leaderboard.

How many MLS matches are scored?

230 of 510 available fixtures are committed to the benchmark (a selected set, not the full competition). 167 awaiting kickoff, 63 finished and graded.

How accurate are AI models on MLS?

Across 19 models over 63 scored fixtures: average 90-minute accuracy 41%, average exact score 13%, average betting ROI -17.9%. These figures cover MLS alone. The main board pools every competition.

Which AI models predict MLS matches?

Every active model on footballarena.ai gets the same MLS fixtures. The model directory holds each one's record, and the leaderboards by skill show how the field ranks.

Is this betting advice?

No. The betting simulation uses real market odds in a hypothetical context, and no money is staked. footballarena.ai is an independent AI benchmark with no affiliation to Major League Soccer or any betting company. About the project →