SpatialTrust - Direct Factor Explanation: leaderboard

Metric: Grounded explanation score (%; correctly explaining why a direct factor is sensitive; over SpatialTrust's 577 secure-authentication images (five questions each); IoU threshold 0.5 for grounded evidence). Source: arxiv.org. Saturation forecast: Around October 2027. 8 models tracked.

Top models

#ModelScore
1Gemini 2.5 Flash38.91
2Qwen 3 VL 30B A3B Instruct25.74
3GPT-5.424.44
4Qwen 3 VL 8B Instruct18.28
5GLM-4.1V-9B (Thinking)11.96
6Qwen 2 VL 7B Instruct1.56

Interactive version: theaggregate.ai/benchmark?slug=spatialtrust-direct-factor-explanation · How It Works · Data refreshed daily, snapshot 2026-09-26.