HCSU - Stone Rubbings: leaderboard

Metric: Accuracy (%) on stone-rubbing (Bei) character images, 8-way controlled candidate selection: identify which of 8 calligraphers (1 correct, 7 distractors, each with a style description and reference image) wrote a target character image, temperature 0.1; chance 12.5; higher is better. Source: arxiv.org. Saturation forecast: Around 2029. 9 models tracked.

Top models

#ModelScore
1GPT-5.235.32
2Claude Sonnet 4.534.26
3Qwen 3 VL 235B A22B Instruct34.1
4GPT-4o32.26
5Gemini 2.5 Pro31.49
6Qwen 2.5 VL 72B Instruct26.6
7Qwen 3 VL 30B A3B Instruct23.42
8Qwen 2.5 VL 7B Instruct21.27

Interactive version: theaggregate.ai/benchmark?slug=hcsu-stone-rubbings · How It Works · Data refreshed daily, snapshot 2026-09-29.