PortBench-QA - Position Sizing: leaderboard
PortBench-QA's position-sizing template (Zhao, Chen and Su, 2026): the model sizes a position with a fixed-fractional rule from a maximum allowed drawdown and the asset's 99% value at risk. Item score in percent over 50 test-period (2024-2025) questions, zero-shot at temperature 0 with thinking on; higher is better.
Metric: Item score (%; fixed-fractional sizing from a drawdown limit and 99% VaR, 50 test questions per model). Source: arxiv.org. Saturation forecast: Estimated already saturated. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.6 Plus (Thinking) | 96.8 | #217 (Qwen 3.6 Plus) |
| 2 | GLM-5.1 (Thinking) | 96.4 | #175 (GLM-5.1) |
| 3 | Qwen 3.6 35B A3B (Thinking) | 96.1 | #461 (Qwen 3.6 35B A3B) |
| 4 | Qwen 3.7 Max (Thinking) | 95.1 | #142 (Qwen 3.7 Max) |
| 5 | Kimi K2.6 (Thinking) | 49.3 | #173 (Kimi K2.6) |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=portbench-qa-position-sizing · How It Works · Data refreshed daily, snapshot 2026-11-12.