ActTraitBench: leaderboard
Metric: G_KD (Global Knowledge-Decision Gap) (self-reported). Source: benchmarklist.com. 15 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | MiniMax-M2.5 | 2.17 |
| 2 | Gemini 3.1 Pro (Preview) | 1.83 |
| 3 | GLM-5 | 1.79 |
| 4 | Qwen 3 235B A22B (Thinking) | 1.54 |
| 5 | Qwen 3 32B | 0.85 |
| 6 | DeepSeek V3.1 | 0.75 |
| 7 | DeepSeek V4 Flash | 0.65 |
| 8 | DeepSeek V3.2 | 0.59 |
| 9 | DeepSeek V4 Pro | 0.57 |
| 10 | Claude Sonnet 4.6 | 0.51 |
| 11 | DeepSeek V3 | 0.47 |
| 12 | GPT-4o | 0.28 |
| 13 | Qwen 3 8B | 0.23 |
| 14 | Qwen 3 1.7B | 0.19 |
Interactive version: theaggregate.ai/benchmark?slug=acttraitbench · How It Works · Data refreshed daily, snapshot 2026-09-05.