🤖 AI benchmark: hit-rate of 7 models

Prematch Live (in-play)

Seven external AI models (Hermes contour) independently analyze the same matches — predicting the outcome (1X2), total (Over/Under), both teams to score (BTTS) and the exact score. Here we honestly compare their predictions against the real result after the final whistle and combine everything into a single accuracy rating. An informational and analytical snapshot, not betting advice.

⚠️ Data is still accumulating — counting starts from 09.07.2026, so all models are compared on the same events (early test predictions are excluded). The sample is still small and not representative. Right now the snapshot holds 119 match(es), 480 settled AI predictions (Volleyball). The figures below are N, not «a percentage you can trust»: the more matches are played out, the more reliable the snapshot becomes. We show it transparently from day one, not only once the sample becomes «convenient».

All sports Football Tennis Basketball Volleyball Hockey Table tennis

Leaderboard · Volleyball

Model	N (settled)	1X2	Exact score	Composite accuracy
Kimi	63	65.1%(41/63)	31.9%(15/47)	50.9%(56/110)
ChatGPT	18	66.7%(12/18)	18.2%(2/11)	48.3%(14/29)
Claude	98	65.3%(64/98)	19.5%(15/77)	45.1%(79/175)
DeepSeek	13	76.9%(10/13)	7.7%(1/13)	42.3%(11/26)
Google AI	124	64.5%(80/124)	19.4%(24/124)	41.9%(104/248)
Qwen	82	69.5%(57/82)	7.9%(6/76)	39.9%(63/158)
GLM 5.2	82	65.9%(54/82)	11.0%(9/82)	38.4%(63/164)

grey — sample <5, not representative; «—» — the model has not made a settled prediction yet.

Composite accuracy — the share of correct predictions across all shown markets together: (sum of correct picks) ÷ (sum of all settled picks) across the markets 1X2 + Exact score. Each market-pick weighs equally. This is hit-rate, not profitability — for money/ROI by model see /ai-agent. «Exact score» — the full final score was guessed correctly (H and A matched); predictions with no recognized score do not count toward the denominator.

Composite model rating · all markets · Volleyball

Bar height = the model's composite accuracy across all applicable markets on the current sample. Sorted from best to worst.

50.9% (56/110)

Kimi 2.6

48.3% (14/29)

GPT 5.5

45.1% (79/175)

Opus 4.8

42.3% (11/26)

DeepSeek V4 Pro

41.9% (104/248)

Gemini 3.5 Flash

39.9% (63/158)

Qwen 3.7 Plus

38.4% (63/164)

GLM 5.2

Bars are AI models by version; grey/dimmed — sample <5, not representative. The snapshot is informational, not betting advice.

Accuracy by market · Volleyball

Where each model is strong: one mini-bar per applicable market, with the percentage and (hits/sample).

Kimi Composite 50.9%

1X2

65.1% (41/63)

Exact score

31.9% (15/47)

ChatGPT Composite 48.3%

1X2

66.7% (12/18)

Exact score

18.2% (2/11)

Claude Composite 45.1%

1X2

65.3% (64/98)

Exact score

19.5% (15/77)

DeepSeek Composite 42.3%

1X2

76.9% (10/13)

Exact score

7.7% (1/13)

Google AI Composite 41.9%

1X2

64.5% (80/124)

Exact score

19.4% (24/124)

Qwen Composite 39.9%

1X2

69.5% (57/82)

Exact score

7.9% (6/76)

GLM 5.2 Composite 38.4%

1X2

65.9% (54/82)

Exact score

11.0% (9/82)

The model's favorite by 1X2 = the max of P1/X/P2 in its probabilities; for sports without a draw (tennis, volleyball, etc.) the «X» option doesn't participate. grey — sample <5, not representative. Not betting advice.