PredictionArena (Kalshi) — leaderboard

Live prediction-market benchmark from Arcada Labs where LLM agents trade Kalshi event markets and are ranked by account value. Tests forecasting, research, risk management, and trading decisions with settlement and liquidity constraints.

Metric: Account Value ($). Source: www.predictionarena.ai. Status: saturation imminent. 10 models tracked.

Top models

#ModelScore
1Claude Opus 4.61062592.74
2GLM-520988.39
3Gemini 3.1 Pro (Preview)17425.7
4GPT-5.410121.56
5GLM-4.78459.6
6Grok 4.207996.41
7Claude Opus 4.5 (20251101)7337.75
8GPT-5.27322.19
9Grok 4.1 Fast (Reasoning)6924.88
10Gemini 3 Pro (Preview)6903.08

Interactive version: theaggregate.ai/benchmark?slug=predictionarena-kalshi · How the rankings work · Data refreshed daily, snapshot 2026-07-22.