Dataset Card for "MM-IQ"

• Dataset Description
• Paper Information
• Dataset Examples
• Leaderboard
• Dataset Usage
  • Data Downloading
  • Data Format
  • Automatic Evaluation
• Citation

Dataset Description

MM-IQ is a new benchmark designed to evaluate MLLMs' intelligence through multiple reasoning patterns demanding abstract reasoning abilities. It encompasses three input formats, six problem configurations, and eight reasoning patterns. With 2,710 samples, MM-IQ is the most comprehensive and largest AVR benchmark for evaluating the intelligence of MLLMs, and 3x and 10x larger than two very recent benchmarks MARVEL and MathVista-IQTest, respectively. By focusing on AVR problems, MM-IQ provides a targeted assessment of the cognitive capabilities and intelligence of MLLMs, contributing to a more comprehensive understanding of their strengths and limitations in the pursuit of AGI.

Paper Information

• Paper: Coming soon.
• Code: https://github.com/AceCHQ/MMIQ/tree/main
• Project: https://acechq.github.io/MMIQ-benchmark/
• Leaderboard: https://acechq.github.io/MMIQ-benchmark/#leaderboard

Dataset Examples

Examples of our MM-IQ:

1. Logical Operation Reasoning

Prompt: Choose the most appropriate option from the given four choices to fill in the question mark, so that it presents a certain regularity:

🔍 Click to expand/collapse more examples

2. Mathematical Reasoning

   Prompt1: Choose the most appropriate option from the given four options to present a certain regularity:

   Option A: 4; Option B: 5; Option C: 6; Option D: 7.

   

3. 2D-geometry Reasoning

   Prompt: The option that best fits the given pattern of figures is ( ).

   

4. 3D-geometry Reasoning

   Prompt: The one that matches the top view is:

   

5. visual instruction Reasoning

   Prompt: Choose the most appropriate option from the given four options to present a certain regularity:

   

6. Spatial Relationship Reasoning

   Prompt: Choose the most appropriate option from the given four options to present a certain regularity:

   

7. Concrete Object Reasoning

   Prompt: Choose the most appropriate option from the given four choices to fill in the question mark, so that it presents a certain regularity:

   

8. Temporal Movement Reasoning

   Prompt:Choose the most appropriate option from the given four choices to fill in the question mark, so that it presents a certain regularity:

   

Leaderboard

🏆 The leaderboard for the MM-IQ (2,710 problems) is available here.

Dataset Usage

Data Downloading

You can download this dataset by the following command (make sure that you have installed Huggingface Datasets):

from IPython.display import display, Image
from datasets import load_dataset

dataset = load_dataset("huanqia/MM-IQ")

Here are some examples of how to access the downloaded dataset:

# print the first example on the MM-IQ dataset
print(dataset["test"][0])
print(dataset["test"][0]['data_id']) # print the problem id 
print(dataset["test"][0]['question']) # print the question text 
print(dataset["test"][0]['answer']) # print the answer
# Display the image
print("Image context:")
display(dataset["test"][0]['image'])

We have uploaded a demo to illustrate how to access the MM-IQ dataset on Hugging Face, available at hugging_face_dataset_demo.ipynb.

Data Format

The dataset is provided in Parquet format and contains the following attributes:

{
    "question": [string] The question text,
    "answer": [string] The correct answer for the problem,
    "data_id": [int] The problem id,
    "category": [string] The category of reasoning pattern,
    "image": [image] Containing image (raw bytes and image path) corresponding to the image in data.zip,
}

Automatic Evaluation

🔔 To automatically evaluate a model on the dataset, please refer to our GitHub repository here.

Citation

If you use the MM-IQ dataset in your work, please kindly cite the paper using this BibTeX:

@misc{cai2025mm-iq,
  title = {MM-IQ: Benchmarking Human-Like Abstraction and Reasoning in Multimodal Models},
  author = {Huanqia Cai and Yijun Yang and Winston Hu},
  month = {January},
  year = {2025}
}

Contact

Huanqia Cai: [email protected]