RuleWeaver - Same-Source - Rule Precision: leaderboard
Metric: Correctly applied share of cited rules (%). Source: arxiv.org. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.5 (Non-reasoning) | 74.58 |
| 2 | GPT-5.4 (Non-reasoning) | 71.31 |
| 3 | Claude Opus 4.6 (Non-reasoning) | 70.41 |
| 4 | GLM-5 (Non-reasoning) | 64.17 |
| 5 | Claude Sonnet 4.6 (Non-reasoning) | 63.15 |
| 6 | DeepSeek V4 Pro (Non-reasoning) | 59.22 |
| 7 | Gemini 3.1 Pro (Preview) (Low) | 58.44 |
| 8 | Kimi K2.6 (Non-reasoning) | 55.13 |
| 9 | MiniMax-M2.7 | 47.43 |
Interactive version: theaggregate.ai/benchmark?slug=ruleweaver-same-source-rule-precision · How It Works · Data refreshed daily, snapshot 2026-09-19.