TurtleAI - Real: leaderboard

Metric: Success rate (%) on the 102 TurtleAI-DSReal tasks (XLogoOnline Midi and Maxi levels designed for grades 3-6), symbolic comparison: the predicted Python turtle program is run in a canonicalizing emulator and succeeds when its normalized drawing matches the reference pixel by pixel (0.95 or 0.92 overlap threshold), greedy decoding, one fixed prompt with the task image; higher is better. Source: arxiv.org. Saturation forecast: Around March 2027. 18 models tracked.

Top models

#ModelScore
1O3 (Medium)40.2
2O4 Mini (Medium)36.27
3GPT-5 (Medium)27.45
4GPT-4o (2024-11-20)26.47
5Qwen 3 VL 30B A3B Instruct18.63
6GPT-4 Turbo15.69
7Qwen 2.5 VL 72B Instruct15.69
8InternVL3-78B12.75
9Pixtral-12B9.8
10Qwen 2 VL 7B Instruct0.98

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