SafeSceneReason: leaderboard
SafeSceneReason (Shanghai Jiao Tong University and collaborators, 2026) is a multimodal industrial-safety reasoning benchmark linking workplace scenes and accident-report figures with multiple-choice questions. Micro accuracy over all 1,000 test questions, in percent; higher is better.
Metric: Accuracy (%; micro average over all 1,000 questions; single-answer multiple choice on the 1,000-question test set balanced across answer positions, built from workplace scene graphs and NIOSH FACE and CSB accident-report figures; temperature 0, unparseable answers counted wrong). Source: arxiv.org. Saturation forecast: Around February 2027. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.6 Sol | 88.7 | #29 |
| 2 | GPT-5.5 | 87.9 | #48 |
| 3 | Gemini 3.1 Pro (Preview) | 80.3 | #91 |
| 4 | Gemini 3.5 Flash | 78.7 | #98 |
| 5 | Claude Fable 5 | 74.1 | #26 |
| 6 | Kimi K2.6 | 70.7 | #173 |
| 7 | Qwen 3.6 27B | 40.9 | #339 |
| 8 | Qwen 3.5 9B | 37.4 | #803 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=safescenereason · How It Works · Data refreshed daily, snapshot 2026-11-12.