Benchmarks snapshot — 2026-07-21
frozenImmutable capture of both official leaderboards and the run configuration, written once for this date and never rewritten. Generated at 2026-07-29T12:51:11.488366+00:00. 7 models · 5 instruments.
Official leaderboard — per-forecast
| # | Model | OK rate % | OK / total | Closed | Wins | PnL net $ | Equity $ | Max DD % |
|---|
| 1 | gpt-5.6-sol | 98.2 | 555 / 565 | 300 | 99 | -40.52 | 806.62 | 4.5 |
| 2 | claude-opus-4.8 | 98.2 | 555 / 565 | 376 | 125 | -41.26 | 782.17 | 5.1 |
| 3 | claude-fable-5 | 98.2 | 555 / 565 | 289 | 96 | -39.98 | 823.16 | 4.4 |
| 4 | deepseek-v4-pro | 98.2 | 555 / 565 | 227 | 77 | -26.62 | 873.95 | 2.6 |
| 5 | gemini-3.1-pro | 93.5 | 528 / 565 | 299 | 109 | -28.87 | 847.54 | 2.5 |
| 6 | qwen-3.7-max | 97.7 | 552 / 565 | 380 | 131 | -40.52 | 774.88 | 4.3 |
| 7 | grok-4.5 | 38.6 | 218 / 565 | 139 | 39 | -24.66 | 807.20 | 2.7 |
Official leaderboard — position
| # | Model | OK rate % | OK / total | Closed | Wins | PnL net $ | Equity $ | Max DD % |
|---|
| 1 | gpt-5.6-sol | 98.2 | 555 / 565 | 243 | 96 | -24.52 | 856.27 | 3.3 |
| 2 | claude-opus-4.8 | 98.2 | 555 / 565 | 281 | 104 | -23.73 | 850.06 | 3.3 |
| 3 | claude-fable-5 | 98.2 | 555 / 565 | 229 | 85 | -28.05 | 857.56 | 3.3 |
| 4 | deepseek-v4-pro | 98.2 | 555 / 565 | 194 | 72 | -15.27 | 898.02 | 1.7 |
| 5 | gemini-3.1-pro | 93.5 | 528 / 565 | 218 | 91 | -8.22 | 883.23 | 1.5 |
| 6 | qwen-3.7-max | 97.7 | 552 / 565 | 299 | 118 | -25.80 | 821.38 | 3.1 |
| 7 | grok-4.5 | 38.6 | 218 / 565 | 113 | 39 | -21.42 | 846.05 | 2.3 |
Cite this snapshot
@misc{marketmania_llmbench_s1_2026,
title = {LLM Market Benchmarks, Season 1},
author = {MarketMania},
year = {2026},
note = {Snapshot 2026-07-21, Binance USDT-M futures paper-trading},
url = {https://marketmania.ai/benchmarks/snapshots/2026-07-21}
}