KellyBench - Strategy Sophistication: leaderboard

Strategy sophistication on KellyBench (2026), a season-long Premier League betting environment: an expert rubric of 45 criteria on feature development, modelling, wagering and handling non-stationarity, written with quantitative betting practitioners, scores each run. Percent of the maximum rubric score; higher is better.

Metric: Strategy sophistication (% of the 52-point expert rubric of 45 criteria on feature development, modelling, wagering and handling non-stationarity, built with practitioners from quantitative betting funds), averaged over five seeds on KellyBench's sequential simulation of the 2023/24 English Premier League season, where the agent builds betting models from historical data, lineups and public odds and places bets matchday by matchday from a 100,000 GBP bankroll (ReSum harness, maximum reasoning budget with interleaved thinking, 500-1000 tool calls per episode), five seeds; higher is better. Source: arxiv.org. Saturation forecast: Around May 2028. 5 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Claude Opus 4.6Max26.5#113
2GPT-5.4xHigh22.3#135
3GLM-5Thinking17.3#269
4Kimi K2.5Thinking12.7#266
5Gemini 3.1 Pro (Preview)8.8#99

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=kellybench-strategy-sophistication · How It Works · Data refreshed daily, snapshot 2026-11-21.