TurtleAI: Benchmarking Multimodal Models for Visual Programming in Turtle Graphics
For researchers and educators, this benchmark reveals current VLMs' limitations in spatial reasoning for educational visual programming and provides a method to improve performance.
TurtleAI benchmarks vision-language models on education-oriented visual programming tasks in Turtle Graphics, finding most models achieve success rates below 30%. Fine-tuning Qwen2-VL-72B on synthetic data improves performance by about 20% on real-world tasks.
Vision-language models (VLMs) have been explored for visual programming, where they generate code to solve visual tasks. However, most prior work focuses on visual programming for productivity; it remains unclear how well current VLMs perform on education-oriented visual programming and what factors limit their performance. To bridge this gap, we introduce TurtleAI, a benchmark containing 823 tasks curated based on real-world visual programming tasks in the Turtle Graphics domain. Solving these tasks requires models to perceive geometric patterns, reason about spatial relationships, and synthesize Python code that faithfully reproduces geometric patterns. We evaluate 20+ VLMs, including GPT-5, GPT-4o, and Qwen2-VL-72B, and find that they struggle significantly, with most achieving success rates below 30%. To address these limitations, we propose a data generation technique that requires only a small set of seed samples. Fine-tuning Qwen2-VL-72B on the resulting synthetic data yields an improvement of about 20% on real-world tasks. Our failure analysis reveals that GPT-4o struggles with spatial reasoning and precise visual replication, whereas fine-tuning primarily improves the alignment between visual reasoning and code implementation.