PortBench-QA - Regime Detection: leaderboard

Metric: Item score (%; market regime plus the implied allocation direction for six asset classes, partial credit, 50 test questions per model). Source: arxiv.org. Saturation forecast: Around December 2026. 10 models tracked.

Top models

#ModelScore
1DeepSeek V4 Flash (Thinking)84.3
2Qwen 3.6 Plus (Thinking)76.8
3Qwen 3.6 35B A3B (Thinking)76.3
4DeepSeek V4 Pro (Thinking)76
5Qwen 3.7 Max (Thinking)74.2
6GLM-5.1 (Thinking)73.8
7Kimi K2.6 (Thinking)32

Interactive version: theaggregate.ai/benchmark?slug=portbench-qa-regime-detection · How It Works · Data refreshed daily, snapshot 2026-09-26.