PaperAudit-Bench (Fast Mode) - Qwen3-VL-235B-A22B Errors: leaderboard
PaperAudit-Bench (Tu et al., 2026) tests whether a model finds errors injected into accepted papers. This board uses the 46 NeurIPS 2025 oral papers as corrupted by Qwen3-VL-235B-A22B (10 to 20 injected errors of eight types each), reviewed by each model in the Fast mode's single pass over the whole paper; text-only models see the papers without figures. Macro-F1 in percent over papers, the harmonic mean of error coverage and finding precision with a GPT-5.1 judge matching findings to injected errors; higher is better.
Metric: Macro-F1 (%; the per-paper harmonic mean of error coverage and finding precision, averaged over the 46 NeurIPS 2025 oral papers of the benchmark's NeurIPS branch as corrupted by qwen3-vl-235b-a22b-instruct with 10-20 injected errors of eight types; Fast mode, one global review pass over the whole paper; findings matched to the injected errors by a GPT-5.1 judge; the three text-only detectors read the papers with figures removed). Source: arxiv.org. Saturation forecast: Around February 2027. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 38.4 | #266 |
| 2 | GPT-5 | 38.1 | #167 |
| 3 | Qwen 3 235B A22B 2507 Instruct | 31.2 | #619 |
| 4 | Grok 4 | 31 | #307 |
| 5 | Claude Sonnet 4.5 | 29.6 | #250 |
| 6 | O4 Mini | 27.4 | #353 |
| 7 | GLM-4.6 (Non-reasoning) | 27 | #486 (GLM-4.6) |
| 8 | DeepSeek V3.1 (Non-reasoning) | 21.7 | #497 (DeepSeek V3.1) |
| 9 | Kimi K2 0905 | 15.4 | #558 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=paperaudit-bench-fast-mode-qwen3-vl-235b-a22b-errors · How It Works · Data refreshed daily, snapshot 2026-11-13.