SafeSceneReason - Quantitative Reasoning: leaderboard

Metric: Accuracy (%; the 279 quantitative reasoning 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 December 2026. 8 models tracked.

Top models

#ModelScore
1GPT-5.6 Sol87.81
2GPT-5.585.3
3Gemini 3.1 Pro (Preview)82.08
4Gemini 3.5 Flash78.85
5Kimi K2.673.84
6Claude Fable 572.04
7Qwen 3.6 27B31.54
8Qwen 3.5 9B24.73

Interactive version: theaggregate.ai/benchmark?slug=safescenereason-quantitative-reasoning · How It Works · Data refreshed daily, snapshot 2026-09-26.