From Pixels to BFS: High Maze Accuracy Does Not Imply Visual Planning
How do multimodal models solve visual spatial tasks -- through genuine planning, or through brute-force search in token space? We introduce \textsc{MazeBench}, a benchmark of 110 procedurally generated maze images across nine controlled groups, and evaluate 16 model configurations from OpenAI, Anthropic, Google, and Alibaba. GPT-5.4 solves 91\% and Gemini 3.1 Pro 79\%, but these scores are misleading: models typically translate images into text grids and then enumerate paths step by step, consuming 1,710--22,818 tokens per solve for a task humans do quickly. Without added reasoning budgets, all configurations score only 2--12\%; on 20$\times$20 ultra-hard mazes, they hit token limits and ...
| Authors | Alberto G. Rodriguez Salgado |
|---|---|
| Primary category | cs.LG |
| Published | 27 March 2026 |
| Last updated at source | 12 May 2026 |
Popularity, last 2 days
0 now. Counts as reported by the source, snapshotted daily by Silytics.
Connected in the knowledge graph
- albertoRodriguez97/MazeBenchDataset · cited by
Models and datasets referencing this paper
- albertoRodriguez97/MazeBenchDataset
Silytics shows public metadata from arXiv and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.