AppellateGen - Reversal Prediction Accuracy: leaderboard
Metric: Ruling-Direction Accuracy (%; affirm vs reverse, balanced 500/500). Source: arxiv.org. Saturation forecast: Around 2028. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Flash (Non-reasoning) | 66.53 |
| 2 | Qwen 3 30B A3B 2507 (Thinking) | 64.81 |
| 3 | Qwen 3 235B A22B 2507 Instruct | 61.42 |
| 4 | Qwen 3 30B A3B 2507 Instruct | 60.43 |
| 5 | QwQ-32B | 60.06 |
| 6 | Qwen 3 8B (Non-reasoning) | 57.92 |
| 7 | DeepSeek R1 0528 Qwen3 8B | 57.87 |
| 8 | Gemini 2.5 Flash Lite (Non-reasoning) | 57.58 |
| 9 | GPT-OSS-120B | 41.81 |
Interactive version: theaggregate.ai/benchmark?slug=appellategen-reversal-prediction-accuracy · How It Works · Data refreshed daily, snapshot 2026-09-25.