WDCD: leaderboard
WDCD — whole-document contradiction detection; composite DCD score across the three rounds.
Metric: DCD Score. Source: huggingface.co. 15 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Grok 4 | 91.97 | #181 |
| 2 | Claude Opus 4.7 | 90.66 | #49 |
| 3 | DeepSeek V4 Pro | 90.38 | #102 |
| 4 | Gemini 3.1 Pro (Preview) | 89.76 | #58 |
| 5 | O3 | 87.07 | #124 |
| 6 | GLM-4.6 | 86.59 | #252 |
| 7 | Claude Sonnet 4.6 | 86.52 | #90 |
| 8 | GPT-6 Sol | 85.31 | #14 |
| 9 | GPT-6 Astra | 82.28 | #6 |
| 10 | GPT-6 Luna | 81.48 | #56 |
| 11 | GPT-5.5 | 80.24 | #31 |
| 12 | GPT-6.1 Sol | 79.9 | #7 |
| 13 | Gemini 2.5 Pro | 78.34 | #152 |
| 14 | Qwen 3 Max | 61.55 | #208 |
Interactive version: theaggregate.ai/benchmark?slug=wdcd · How It Works · Data refreshed daily, snapshot 2026-10-20.