CurveBench - Hard: leaderboard

Metric: Tree-Generation Accuracy (%; exact containment tree, 456 polygon, topographic, maze and counting images, one rollout). Source: arxiv.org. Saturation forecast: Around September 2028. 9 models tracked.

Top models

#ModelScore
1Gemini 3.1 Pro (Preview)19.1
2GPT-5.46.6
3Claude Opus 4.54.2
4Qwen 3 VL 8B (Thinking)4.2
5GPT-5.4 Mini2.4
6Gemma 3 12B (IT)0.7

Interactive version: theaggregate.ai/benchmark?slug=curvebench-hard · How It Works · Data refreshed daily, snapshot 2026-09-25.