PaperAudit-Bench - Section-Level Detection@1: leaderboard
Metric: F1 (%; harmonic mean of error coverage and finding precision on 1,209 corrupted sections of the ICML branch, the model prompted to report only its single most confident error; Qwen3-80B matching judge). Source: arxiv.org. Saturation forecast: Around December 2026. 7 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Pro | 50.6 |
| 2 | Claude Sonnet 4.5 | 48.3 |
| 3 | Qwen 3 235B A22B | 45.9 |
| 4 | DeepSeek V3.1 | 30.6 |
| 5 | Qwen 3 14B | 18.5 |
| 6 | Qwen 3 8B | 16.6 |
Interactive version: theaggregate.ai/benchmark?slug=paperaudit-bench-section-level-detection-1 · How It Works · Data refreshed daily, snapshot 2026-09-26.