# How Well Can Vison-Language Models Understand Humans' Intention? An Open-ended Theory of Mind Question Evaluation Benchmark

arXiv: https://arxiv.org/abs/2503.22093 · HTML: https://mlsanity.com/paper/2503.22093/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Multimodal reasoning, one of 17 directions in LLM reasoning.
- Builds on 7 of its 13 references that are in the map.
- First posted on arXiv 28 Mar 2025.

## Placed in

- LLM reasoning / [Multimodal reasoning](https://mlsanity.com/lr/multimodal-reasoning/)
- Multimodal LLMs / vision-language models / [VLM benchmarks and visual reasoning](https://mlsanity.com/mllm/vlm-benchmarks-and-visual-reasoning/)

## Builds on (most cited first, 7 of 7)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [Visual Instruction Tuning](https://mlsanity.com/paper/2304.08485/) (2023)
- [DeepSeek LLM: Scaling Open-Source Language Models with Longtermism](https://mlsanity.com/paper/2401.02954/) (2024)
- [Do Androids Laugh at Electric Sheep? Humor "Understanding" Benchmarks from The New Yorker Caption Contest](https://mlsanity.com/paper/2209.06293/) (2022)
- [MMToM-QA: Multimodal Theory of Mind Question Answering](https://mlsanity.com/paper/2401.08743/) (2024)
- [Zero-Shot Visual Reasoning by Vision-Language Models: Benchmarking and Analysis](https://mlsanity.com/paper/2409.00106/) (2024)
- [Emotional Theory of Mind: Bridging Fast Visual Processing with Slow Linguistic Reasoning](https://mlsanity.com/paper/2310.19995/) (2023)

## Built on by (most cited first, 1 of 1)

- [Are Vision Language Models Cross-Cultural Theory of Mind Reasoners?](https://mlsanity.com/paper/2512.17394/) (2025)

## Abstract (arXiv)

Vision Language Models (VLMs) have demonstrated strong reasoning capabilities in Visual Question Answering (VQA) tasks; however, their ability to perform Theory of Mind (ToM) tasks, such as inferring human intentions, beliefs, and mental states, remains underexplored. We propose an open-ended question framework to evaluate VLMs' performance across diverse categories of ToM tasks. We curated and annotated a benchmark dataset of 30 images and evaluated the performance of four VLMs of varying sizes. Our results show that the GPT-4 model outperformed all the others, with only one smaller model, GPT-4o-mini, achieving comparable performance. We observed that VLMs often struggle to infer intentions in complex scenarios such as bullying or cheating. Our findings reveal that smaller models can sometimes infer correct intentions despite relying on incorrect visual cues. The dataset is available at https://github.com/ximingwen/ToM-AAAI25-Multimodal.