LogicScore - 2WikiMultiHopQA Determinateness: leaderboard
Metric: Determinateness (%; answers whose own short answer o4-mini re-derives from the long-form answer alone; 2,177 2WikiMultiHopQA items). Source: arxiv.org. Saturation forecast: Estimated already saturated. 21 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3 235B A22B | 96.74 |
| 2 | O3 | 95.8 |
| 3 | Gemini 3 Pro | 94.69 |
| 4 | GPT-4o | 93.13 |
| 5 | DeepSeek R1 | 93.13 |
| 6 | GPT-5.1 | 92.45 |
| 7 | Qwen 3 30B A3B | 91.86 |
| 8 | Qwen 3 4B | 91.23 |
| 9 | Qwen 3 8B | 90.95 |
| 10 | DeepSeek V3.2 | 90.89 |
| 11 | Qwen 3 14B | 90.55 |
| 12 | Qwen 3 1.7B | 82.53 |
| 13 | Qwen 3 0.6B | 64.03 |
Interactive version: theaggregate.ai/benchmark?slug=logicscore-2wikimultihopqa-determinateness · How It Works · Data refreshed daily, snapshot 2026-09-26.