MapEval: A Map-Based Evaluation of Geo-Spatial Reasoning in Foundation Models

Mahir Labib Dihan, Md Tanvir Hassan, Md Tanvir Parvez, Md Hasebul Hasan, Md Almash Alam, Muhammad Aamir Cheema, Mohammed Eunus Ali, Md Rizwan Parvez
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:13774-13813, 2025.

Abstract

Recent advancements in foundation models have improved autonomous tool usage and reasoning, but their capabilities in map-based reasoning remain underexplored. To address this, we introduce MapEval, a benchmark designed to assess foundation models across three distinct tasks—textual, API-based, and visual reasoning—through 700 multiple-choice questions spanning 180 cities and 54 countries, covering spatial relationships, navigation, travel planning, and real-world map interactions. Unlike prior benchmarks that focus on simple location queries, MapEval requires models to handle long-context reasoning, API interactions and visual map analysis, making it the most comprehensive evaluation framework for geospatial AI. On evaluation of 30 foundation models, including Claude-3.5-Sonnet, GPT-4o, Gemini-1.5-Pro, none surpasses 67% accuracy, with open-source models performing significantly worse and all models lagging over 20% behind human performance. These results expose critical gaps in spatial inference, as models struggle with distances, directions, route planning, and place-specific reasoning, highlighting the need for better geospatial AI to bridge the gap between foundation models and real-world navigation.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-dihan25a, title = {{M}ap{E}val: A Map-Based Evaluation of Geo-Spatial Reasoning in Foundation Models}, author = {Dihan, Mahir Labib and Hassan, Md Tanvir and Parvez, Md Tanvir and Hasan, Md Hasebul and Alam, Md Almash and Cheema, Muhammad Aamir and Ali, Mohammed Eunus and Parvez, Md Rizwan}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {13774--13813}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/dihan25a/dihan25a.pdf}, url = {https://proceedings.mlr.press/v267/dihan25a.html}, abstract = {Recent advancements in foundation models have improved autonomous tool usage and reasoning, but their capabilities in map-based reasoning remain underexplored. To address this, we introduce MapEval, a benchmark designed to assess foundation models across three distinct tasks—textual, API-based, and visual reasoning—through 700 multiple-choice questions spanning 180 cities and 54 countries, covering spatial relationships, navigation, travel planning, and real-world map interactions. Unlike prior benchmarks that focus on simple location queries, MapEval requires models to handle long-context reasoning, API interactions and visual map analysis, making it the most comprehensive evaluation framework for geospatial AI. On evaluation of 30 foundation models, including Claude-3.5-Sonnet, GPT-4o, Gemini-1.5-Pro, none surpasses 67% accuracy, with open-source models performing significantly worse and all models lagging over 20% behind human performance. These results expose critical gaps in spatial inference, as models struggle with distances, directions, route planning, and place-specific reasoning, highlighting the need for better geospatial AI to bridge the gap between foundation models and real-world navigation.} }
Endnote
%0 Conference Paper %T MapEval: A Map-Based Evaluation of Geo-Spatial Reasoning in Foundation Models %A Mahir Labib Dihan %A Md Tanvir Hassan %A Md Tanvir Parvez %A Md Hasebul Hasan %A Md Almash Alam %A Muhammad Aamir Cheema %A Mohammed Eunus Ali %A Md Rizwan Parvez %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-dihan25a %I PMLR %P 13774--13813 %U https://proceedings.mlr.press/v267/dihan25a.html %V 267 %X Recent advancements in foundation models have improved autonomous tool usage and reasoning, but their capabilities in map-based reasoning remain underexplored. To address this, we introduce MapEval, a benchmark designed to assess foundation models across three distinct tasks—textual, API-based, and visual reasoning—through 700 multiple-choice questions spanning 180 cities and 54 countries, covering spatial relationships, navigation, travel planning, and real-world map interactions. Unlike prior benchmarks that focus on simple location queries, MapEval requires models to handle long-context reasoning, API interactions and visual map analysis, making it the most comprehensive evaluation framework for geospatial AI. On evaluation of 30 foundation models, including Claude-3.5-Sonnet, GPT-4o, Gemini-1.5-Pro, none surpasses 67% accuracy, with open-source models performing significantly worse and all models lagging over 20% behind human performance. These results expose critical gaps in spatial inference, as models struggle with distances, directions, route planning, and place-specific reasoning, highlighting the need for better geospatial AI to bridge the gap between foundation models and real-world navigation.
APA
Dihan, M.L., Hassan, M.T., Parvez, M.T., Hasan, M.H., Alam, M.A., Cheema, M.A., Ali, M.E. & Parvez, M.R.. (2025). MapEval: A Map-Based Evaluation of Geo-Spatial Reasoning in Foundation Models. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:13774-13813 Available from https://proceedings.mlr.press/v267/dihan25a.html.

Related Material