MSA at ImageCLEF 2025 Multimodal Reasoning: Multilingual Multimodal Reasoning With Ensemble Vision Language Models

TL;DR

This paper introduces a robust ensemble system for multilingual multimodal reasoning that combines vision-language models and prompt engineering, achieving top accuracy in the ImageCLEF 2025 challenge across multiple languages.

Contribution

The paper presents a novel ensemble approach integrating multiple vision-language models with prompt strategies, demonstrating superior performance in multilingual multimodal reasoning tasks.

Findings

Ensemble system achieved 81.4% accuracy on the leaderboard.

Prompt design significantly improved model accuracy from 55.9% to 61.7%.

Zero-shot model outperformed trained models in experiments.

Abstract

We present a robust ensemble-based system for multilingual multimodal reasoning, designed for the ImageCLEF 2025 EXAMS V challenge. Our approach integrates Gemini 2.5 Flash for visual description, Gemini 1.5 Pro for caption refinement and consistency checks, and Gemini 2.5 Pro as a reasoner which handles final answer selection, all coordinated through carefully engineered few-shot and zero-shot prompts. We conducted an extensive ablation study, training several large language models (Gemini 2.5 Flash, Phi 4, Gemma 3, Mistral) on an English dataset and its multilingual augmented version. Additionally, we evaluated Gemini 2.5 Flash in a zero-shot setting for comparison and found it to substantially outperform the trained models. Prompt design also proved critical: enforcing concise, language-normalized formats and prohibiting explanatory text boosted model accuracy on the English validation set from 55.9% to 61.7%. On the official leaderboard, our system (Team MSA) achieved first place overall in the multilingual track with 81.4% accuracy, and led 11 out of 13 individual language tracks, with top results such as 95.07% for Croatian and 92.12% for Italian. These findings highlight that lightweight OCR-VLM ensembles, when paired with precise prompt strategies and cross-lingual augmentation, can outperform heavier end-to-end models in high-stakes, multilingual educational settings.