Research map / Video / Video-language understanding
VideoMathQA: Benchmarking Mathematical Reasoning via Multimodal Understanding in Video
Hanoona Abdul Rasheed, Abdelrahman M Shaker, Anqi Tang, Muhammad Maaz, Ming-Hsuan Yang, Salman Khan, Fahad Shahbaz Khan
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add subtitles to video input (ablation) | GPT-o4-mini · VideoMathQA MBin CoT | accuracy: 42.1 -> 44.8 | Improved |
| add subtitles to video input (ablation) | Qwen2.5-VL-72B · VideoMathQA MBin CoT | accuracy: 24.5 -> 28.6 | Improved |
| add subtitles to video input (ablation) | small models (<5B, <9B) · VideoMathQA | accuracy | No clear effect |
| scale InternVL3 from 8B to 38B to 72B (ablation) | InternVL3 · VideoMathQA MBin CoT +Sub | accuracy: 20.0 -> 25.0 -> 27.9 | Improved |
| use newer smaller model instead of older larger model | InternVL3-38B vs LLaVA-Video-72B / LLaVA-OV-72B · VideoMathQA | accuracy | Improved |
| increase input frames from 16 to 768 (ablation) | Qwen2.5-VL · VideoMathQA MCQ CoT | accuracy: up to 8 points on long videos | Improved |
| add chain-of-thought prompting (ablation) | proprietary models · VideoMathQA | accuracy | Improved |
| add chain-of-thought prompting (ablation) | open-source models · VideoMathQA | accuracy | Mixed |
| increase video length from medium to long | most models · VideoMathQA | accuracy | Hurt |
| raise question difficulty from easy to hard | GPT-4o · VideoMathQA | accuracy: 96% -> 46% | Hurt |
Closest papers by meaning
- Benchmarking Multimodal Mathematical Reasoning with Explicit Visual Dependency (NeurIPS 2026 · similarity 0.92)
- MV-MATH: Evaluating Multimodal Math Reasoning in Multi-Visual Contexts (CVPR 2025 · similarity 0.91)
- Measuring Multimodal Mathematical Reasoning with MATH-Vision Dataset (NeurIPS 2024 · similarity 0.90)
- SciVideoBench: Benchmarking Scientific Video Reasoning in Large Multimodal Models (ICML 2026 · similarity 0.90)
- Thinking with Video: Video Generation as a Promising Multimodal Reasoning Paradigm (CVPR 2026 · similarity 0.90)
- VisioMath: Benchmarking Figure-based Mathematical Reasoning in LMMs (ICLR 2026 · similarity 0.90)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
