RotBench: Evaluating Multimodal Large Language Models on Identifying Image Rotation

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Niu, Tianyi, Cho, Jaemin, Stengel-Eskin, Elias, Bansal, Mohit
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917219955376128
author Niu, Tianyi
Cho, Jaemin
Stengel-Eskin, Elias
Bansal, Mohit
author_facet Niu, Tianyi
Cho, Jaemin
Stengel-Eskin, Elias
Bansal, Mohit
contents We investigate to what extent Multimodal Large Language Models (MLLMs) can accurately identify the orientation of input images rotated 0°, 90°, 180°, and 270°. This task demands robust visual reasoning capabilities to detect rotational cues and contextualize spatial relationships within images, regardless of their orientation. To evaluate MLLMs on these abilities, we introduce RotBench, a 350-image manually-filtered benchmark comprising lifestyle, portrait, and landscape images. Despite the relatively simple nature of this task, we show that several state-of-the-art open and proprietary MLLMs, including GPT-5, o3, and Gemini-2.5-Pro, do not reliably identify rotation in input images. Providing models with auxiliary information -- including captions, depth maps, and more -- or using chain-of-thought prompting offers only small and inconsistent improvements. Our results indicate that most models are able to reliably identify right-side-up (0°) images, while certain models are able to identify upside-down (180°) images. None can reliably distinguish between 90° and 270° rotated images. Simultaneously showing the image rotated in different orientations leads to moderate performance gains for reasoning models, while a modified setup using voting improves the performance of weaker models. We further show that fine-tuning does not improve models' ability to distinguish 90° and 270° rotations, despite substantially improving the identification of 180° images. Together, these results reveal a significant gap between MLLMs' spatial reasoning capabilities and human perception in identifying rotation.
format Preprint
id arxiv_https___arxiv_org_abs_2508_13968
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle RotBench: Evaluating Multimodal Large Language Models on Identifying Image Rotation
Niu, Tianyi
Cho, Jaemin
Stengel-Eskin, Elias
Bansal, Mohit
Computer Vision and Pattern Recognition
Artificial Intelligence
Computation and Language
We investigate to what extent Multimodal Large Language Models (MLLMs) can accurately identify the orientation of input images rotated 0°, 90°, 180°, and 270°. This task demands robust visual reasoning capabilities to detect rotational cues and contextualize spatial relationships within images, regardless of their orientation. To evaluate MLLMs on these abilities, we introduce RotBench, a 350-image manually-filtered benchmark comprising lifestyle, portrait, and landscape images. Despite the relatively simple nature of this task, we show that several state-of-the-art open and proprietary MLLMs, including GPT-5, o3, and Gemini-2.5-Pro, do not reliably identify rotation in input images. Providing models with auxiliary information -- including captions, depth maps, and more -- or using chain-of-thought prompting offers only small and inconsistent improvements. Our results indicate that most models are able to reliably identify right-side-up (0°) images, while certain models are able to identify upside-down (180°) images. None can reliably distinguish between 90° and 270° rotated images. Simultaneously showing the image rotated in different orientations leads to moderate performance gains for reasoning models, while a modified setup using voting improves the performance of weaker models. We further show that fine-tuning does not improve models' ability to distinguish 90° and 270° rotations, despite substantially improving the identification of 180° images. Together, these results reveal a significant gap between MLLMs' spatial reasoning capabilities and human perception in identifying rotation.
title RotBench: Evaluating Multimodal Large Language Models on Identifying Image Rotation
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2508.13968