Datasets:
metadata
pretty_name: HighSchoolBench-MCQ
language:
- en
language_creators:
- found
multilinguality:
- monolingual
source_datasets:
- original
task_categories:
- multiple-choice
- visual-question-answering
- question-answering
task_ids:
- multiple-choice-qa
size_categories:
- 1K<n<10K
tags:
- highschoolbench
- benchmark
- education
- high-school
- exam
- stem
- science
- mathematics
- multimodal
- image
- text
- multiple-choice
configs:
- config_name: context_text_images__text_options
default: true
data_files:
- split: train
path: data/context_text_images__text_options/train-*.parquet
- split: validation
path: data/context_text_images__text_options/validation-*.parquet
- split: test
path: data/context_text_images__text_options/test-*.parquet
- config_name: context_text_contextualized_images__image_options
data_files:
- split: test
path: data/context_text_contextualized_images__image_options/test-*.parquet
- config_name: context_text_contextualized_images__text_options
data_files:
- split: train
path: data/context_text_contextualized_images__text_options/train-*.parquet
- split: validation
path: >-
data/context_text_contextualized_images__text_options/validation-*.parquet
- split: test
path: data/context_text_contextualized_images__text_options/test-*.parquet
- config_name: context_text_contextualized_tables__image_options
data_files:
- split: test
path: data/context_text_contextualized_tables__image_options/test-*.parquet
- config_name: context_text_contextualized_tables__text_options
data_files:
- split: train
path: data/context_text_contextualized_tables__text_options/train-*.parquet
- split: validation
path: >-
data/context_text_contextualized_tables__text_options/validation-*.parquet
- split: test
path: data/context_text_contextualized_tables__text_options/test-*.parquet
- config_name: context_text_contextualized_tables_images__image_options
data_files:
- split: test
path: >-
data/context_text_contextualized_tables_images__image_options/test-*.parquet
- config_name: context_text_contextualized_tables_images__text_options
data_files:
- split: train
path: >-
data/context_text_contextualized_tables_images__text_options/train-*.parquet
- split: validation
path: >-
data/context_text_contextualized_tables_images__text_options/validation-*.parquet
- split: test
path: >-
data/context_text_contextualized_tables_images__text_options/test-*.parquet
- config_name: context_text_images__image_options
data_files:
- split: train
path: data/context_text_images__image_options/train-*.parquet
- split: validation
path: data/context_text_images__image_options/validation-*.parquet
- split: test
path: data/context_text_images__image_options/test-*.parquet
- config_name: context_text_only__image_options
data_files:
- split: train
path: data/context_text_only__image_options/train-*.parquet
- split: validation
path: data/context_text_only__image_options/validation-*.parquet
- split: test
path: data/context_text_only__image_options/test-*.parquet
- config_name: context_text_only__table_options
data_files:
- split: train
path: data/context_text_only__table_options/train-*.parquet
- split: validation
path: data/context_text_only__table_options/validation-*.parquet
- split: test
path: data/context_text_only__table_options/test-*.parquet
- config_name: context_text_only__text_options
data_files:
- split: train
path: data/context_text_only__text_options/train-*.parquet
- split: validation
path: data/context_text_only__text_options/validation-*.parquet
- split: test
path: data/context_text_only__text_options/test-*.parquet
- config_name: context_text_table__image_options
data_files:
- split: test
path: data/context_text_table__image_options/test-*.parquet
- config_name: context_text_table__text_options
data_files:
- split: train
path: data/context_text_table__text_options/train-*.parquet
- split: validation
path: data/context_text_table__text_options/validation-*.parquet
- split: test
path: data/context_text_table__text_options/test-*.parquet
- config_name: context_text_table_image__image_options
data_files:
- split: test
path: data/context_text_table_image__image_options/test-*.parquet
- config_name: context_text_table_image__text_options
data_files:
- split: train
path: data/context_text_table_image__text_options/train-*.parquet
- split: validation
path: data/context_text_table_image__text_options/validation-*.parquet
- split: test
path: data/context_text_table_image__text_options/test-*.parquet
dataset_info:
- config_name: context_text_images__text_options
features:
- name: id
dtype: int64
- name: assessment_program
dtype: string
- name: assessment_stage
dtype: string
- name: subject
dtype: string
- name: domain
dtype: string
- name: context
dtype: string
- name: image_context
dtype: string
- name: images
list: image
- name: question
dtype: string
- name: options
list: string
- name: option_labels
list: string
- name: answer
list: string
- name: split
dtype: string
- name: num_answer_options
dtype: int32
- config_name: context_text_contextualized_images__image_options
features:
- name: id
dtype: int64
- name: assessment_program
dtype: string
- name: assessment_stage
dtype: string
- name: subject
dtype: string
- name: domain
dtype: string
- name: context
dtype: string
- name: image_context_1
dtype: string
- name: image_1
dtype: image
- name: image_context_2
dtype: string
- name: image_2
dtype: image
- name: image_context_3
dtype: string
- name: image_3
dtype: image
- name: question
dtype: string
- name: options
list: image
- name: option_labels
list: string
- name: answer
list: string
- name: split
dtype: string
- name: num_answer_options
dtype: int32
- name: option_image_label_groups
list:
list: string
- config_name: context_text_contextualized_images__text_options
features:
- name: id
dtype: int64
- name: assessment_program
dtype: string
- name: assessment_stage
dtype: string
- name: subject
dtype: string
- name: domain
dtype: string
- name: context
dtype: string
- name: image_context_1
dtype: string
- name: image_1
dtype: image
- name: image_context_2
dtype: string
- name: image_2
dtype: image
- name: image_context_3
dtype: string
- name: image_3
dtype: image
- name: image_context_4
dtype: string
- name: image_4
dtype: image
- name: question
dtype: string
- name: options
list: string
- name: option_labels
list: string
- name: answer
list: string
- name: split
dtype: string
- name: num_answer_options
dtype: int32
- config_name: context_text_contextualized_tables__image_options
features:
- name: id
dtype: int64
- name: assessment_program
dtype: string
- name: assessment_stage
dtype: string
- name: subject
dtype: string
- name: domain
dtype: string
- name: context
dtype: string
- name: table_context_1
dtype: string
- name: table_1
dtype: string
- name: table_context_2
dtype: string
- name: table_2
dtype: string
- name: question
dtype: string
- name: options
list: image
- name: option_labels
list: string
- name: answer
list: string
- name: split
dtype: string
- name: num_answer_options
dtype: int32
- name: option_image_label_groups
list:
list: string
- config_name: context_text_contextualized_tables__text_options
features:
- name: id
dtype: int64
- name: assessment_program
dtype: string
- name: assessment_stage
dtype: string
- name: subject
dtype: string
- name: domain
dtype: string
- name: context
dtype: string
- name: table_context_1
dtype: string
- name: table_1
dtype: string
- name: table_context_2
dtype: string
- name: table_2
dtype: string
- name: table_context_3
dtype: string
- name: table_3
dtype: string
- name: table_context_4
dtype: string
- name: table_4
dtype: string
- name: question
dtype: string
- name: options
list: string
- name: option_labels
list: string
- name: answer
list: string
- name: split
dtype: string
- name: num_answer_options
dtype: int32
- config_name: context_text_contextualized_tables_images__image_options
features:
- name: id
dtype: int64
- name: assessment_program
dtype: string
- name: assessment_stage
dtype: string
- name: subject
dtype: string
- name: domain
dtype: string
- name: image_context_1
dtype: string
- name: image_1
dtype: image
- name: image_context_2
dtype: string
- name: image_2
dtype: image
- name: image_context_3
dtype: string
- name: image_3
dtype: image
- name: table_context_1
dtype: string
- name: table_1
dtype: string
- name: table_context_2
dtype: string
- name: table_2
dtype: string
- name: question
dtype: string
- name: options
list: image
- name: option_labels
list: string
- name: answer
list: string
- name: split
dtype: string
- name: num_answer_options
dtype: int32
- name: option_image_label_groups
list:
list: string
- config_name: context_text_contextualized_tables_images__text_options
features:
- name: id
dtype: int64
- name: assessment_program
dtype: string
- name: assessment_stage
dtype: string
- name: subject
dtype: string
- name: domain
dtype: string
- name: context
dtype: string
- name: image_context_1
dtype: string
- name: image_1
dtype: image
- name: image_context_2
dtype: string
- name: image_2
dtype: image
- name: image_context_3
dtype: string
- name: image_3
dtype: image
- name: table_context_1
dtype: string
- name: table_1
dtype: string
- name: table_context_2
dtype: string
- name: table_2
dtype: string
- name: table_context_3
dtype: string
- name: table_3
dtype: string
- name: question
dtype: string
- name: options
list: string
- name: option_labels
list: string
- name: answer
list: string
- name: split
dtype: string
- name: num_answer_options
dtype: int32
- config_name: context_text_images__image_options
features:
- name: id
dtype: int64
- name: assessment_program
dtype: string
- name: assessment_stage
dtype: string
- name: subject
dtype: string
- name: domain
dtype: string
- name: context
dtype: string
- name: images
list: image
- name: question
dtype: string
- name: options
list: image
- name: option_labels
list: string
- name: answer
list: string
- name: split
dtype: string
- name: num_answer_options
dtype: int32
- name: option_image_label_groups
list:
list: string
- config_name: context_text_only__image_options
features:
- name: id
dtype: int64
- name: assessment_program
dtype: string
- name: assessment_stage
dtype: string
- name: subject
dtype: string
- name: domain
dtype: string
- name: context
dtype: string
- name: question
dtype: string
- name: options
list: image
- name: option_labels
list: string
- name: answer
list: string
- name: split
dtype: string
- name: num_answer_options
dtype: int32
- name: option_image_label_groups
list:
list: string
- config_name: context_text_only__table_options
features:
- name: id
dtype: int64
- name: assessment_program
dtype: string
- name: assessment_stage
dtype: string
- name: subject
dtype: string
- name: domain
dtype: string
- name: context
dtype: string
- name: question
dtype: string
- name: options
list: string
- name: option_labels
list: string
- name: answer
list: string
- name: split
dtype: string
- name: num_answer_options
dtype: int32
- config_name: context_text_only__text_options
features:
- name: id
dtype: int64
- name: assessment_program
dtype: string
- name: assessment_stage
dtype: string
- name: subject
dtype: string
- name: domain
dtype: string
- name: context
dtype: string
- name: question
dtype: string
- name: options
list: string
- name: option_labels
list: string
- name: answer
list: string
- name: split
dtype: string
- name: num_answer_options
dtype: int32
- config_name: context_text_table__image_options
features:
- name: id
dtype: int64
- name: assessment_program
dtype: string
- name: assessment_stage
dtype: string
- name: subject
dtype: string
- name: domain
dtype: string
- name: context
dtype: string
- name: tables
dtype: string
- name: question
dtype: string
- name: options
list: image
- name: option_labels
list: string
- name: answer
list: string
- name: split
dtype: string
- name: num_answer_options
dtype: int32
- name: option_image_label_groups
list:
list: string
- config_name: context_text_table__text_options
features:
- name: id
dtype: int64
- name: assessment_program
dtype: string
- name: assessment_stage
dtype: string
- name: subject
dtype: string
- name: domain
dtype: string
- name: context
dtype: string
- name: table_context
dtype: string
- name: tables
dtype: string
- name: question
dtype: string
- name: options
list: string
- name: option_labels
list: string
- name: answer
list: string
- name: split
dtype: string
- name: num_answer_options
dtype: int32
- config_name: context_text_table_image__image_options
features:
- name: id
dtype: int64
- name: assessment_program
dtype: string
- name: assessment_stage
dtype: string
- name: subject
dtype: string
- name: domain
dtype: string
- name: context
dtype: string
- name: images
list: image
- name: tables
dtype: string
- name: question
dtype: string
- name: options
list: image
- name: option_labels
list: string
- name: answer
list: string
- name: split
dtype: string
- name: num_answer_options
dtype: int32
- name: option_image_label_groups
list:
list: string
- config_name: context_text_table_image__text_options
features:
- name: id
dtype: int64
- name: assessment_program
dtype: string
- name: assessment_stage
dtype: string
- name: subject
dtype: string
- name: domain
dtype: string
- name: context
dtype: string
- name: images
list: image
- name: tables
dtype: string
- name: question
dtype: string
- name: options
list: string
- name: option_labels
list: string
- name: answer
list: string
- name: split
dtype: string
- name: num_answer_options
dtype: int32
HighSchoolBench-MCQ
Dataset description
HighSchoolBench-MCQ contains 7,549 multiple-choice questions from high-school standardized assessments and science competitions. It is the multiple-choice component of HighSchoolBench, a multimodal benchmark for evaluating language and vision-language models on high-school assessment questions.
The questions come from assessments such as NAEP, SAT, PSAT, PISA, ACT, and IB, U.S. state examinations, science bowls, and U.S. science olympiads. They include text-based questions and questions involving diagrams, images, and tables. Answer options are provided as text, images, or tables.
Each question includes the correct answer label(s) and metadata identifying its source assessment, subject, and domain. The open-ended component is available separately as HighSchoolBench-OEQ.
Dataset statistics
The dataset contains 7,549 questions and 2,419 images.
| Split | Questions |
|---|---|
train |
1,143 |
validation |
375 |
test |
6,031 |
| Total | 7,549 |
The source split val is exposed as validation on Hugging Face. The split field retains the original source label.
Configurations
The dataset is organized into 15 configurations according to the question context and answer-option format.
| Configuration | Question context | Option format |
|---|---|---|
context_text_only__text_options |
Text | Text |
context_text_only__image_options |
Text | Images |
context_text_only__table_options |
Text | Tables |
context_text_images__text_options |
Text and images | Text |
context_text_images__image_options |
Text and images | Images |
context_text_table__text_options |
Text and tables | Text |
context_text_table__image_options |
Text and tables | Images |
context_text_table_image__text_options |
Text, tables, and images | Text |
context_text_table_image__image_options |
Text, tables, and images | Images |
context_text_contextualized_images__text_options |
Images with associated passages | Text |
context_text_contextualized_images__image_options |
Images with associated passages | Images |
context_text_contextualized_tables__text_options |
Tables with associated passages | Text |
context_text_contextualized_tables__image_options |
Tables with associated passages | Images |
context_text_contextualized_tables_images__text_options |
Tables and images with associated passages | Text |
context_text_contextualized_tables_images__image_options |
Tables and images with associated passages | Images |
In contextualized configurations, individual images and tables are stored alongside their associated passages. Some configurations contain only a test split.
Data format
Main fields
| Field | Description |
|---|---|
id |
Question identifier from the source file. |
assessment_program |
Source assessment or competition. |
assessment_stage |
Assessment stage or exam type. |
subject |
Subject assigned to the question. |
domain |
Topic within the subject. |
context |
Supporting passage or shared context, when available. |
question |
Question stem. Mathematical expressions use LaTeX. |
options |
Ordered answer options as strings or images. Table options are represented as Markdown strings. |
option_labels |
Labels corresponding to the answer options. |
answer |
List of correct option labels. |
split |
Original split label: train, val, or test. |
num_answer_options |
Number of correct options, equal to len(answer). |
Context-specific fields
The following fields appear where relevant to the configuration:
| Field | Description |
|---|---|
images |
List of images accompanying the question. |
image_context |
Text associated with the images. |
tables |
Tables represented as Markdown strings. |
table_context |
Text associated with the tables. |
image_N, image_context_N |
Individual image and its associated passage. |
table_N, table_context_N |
Individual table and its associated passage. |
option_image_label_groups |
Option labels represented in each answer-option image. |
An answer-option image may contain several choices. For example, [["A", "B", "C", "D"]] indicates that one image contains all four labeled choices.
Available files
The dataset is provided in two formats:
- Parquet:
data/<configuration>/, with typed columns and embedded images. These files are used by the Hugging Face dataset viewer andload_dataset. - JSON Lines and images:
subsets/<configuration>/<split>/, containing ametadata.jsonlfile and the images referenced by its records.
In Parquet, options and option_labels are separate ordered lists. In the JSON Lines files, options retain their source representation, usually a mapping such as {"A": "...", "B": "..."}.
Usage
Load a configuration using the Hugging Face datasets library:
from datasets import load_dataset
dataset = load_dataset(
"SciKnowOrg/HighSchoolBench-MCQ",
"context_text_images__text_options",
)
print(dataset)
Access a question and its text answer options:
example = dataset["test"][0]
print("Context:", example["context"])
print("Question:", example["question"])
for label, option in zip(
example["option_labels"],
example["options"],
):
print(f"{label}. {option}")
print("Correct answer:", example["answer"])
# Question images are decoded as PIL images.
images = example["images"]
List the available configurations:
from datasets import get_dataset_config_names
print(get_dataset_config_names("SciKnowOrg/HighSchoolBench-MCQ"))
To download the JSON Lines files and accompanying images:
hf download SciKnowOrg/HighSchoolBench-MCQ \
--repo-type dataset \
--include "subsets/*" \
--local-dir HighSchoolBench-MCQ
Citation
If you use HighSchoolBench, please cite:
@misc{rani2026highschoolbench,
title = {HighSchoolBench: A Multimodal Benchmark of High School Assessment Questions},
author = {Rani, Krishna and D'Souza, Jennifer},
year = {2026},
publisher = {Hugging Face},
howpublished = {\url{https://huggingface.co/datasets/SciKnowOrg/HighSchoolBench-MCQ},
\url{https://huggingface.co/datasets/SciKnowOrg/HighSchoolBench-OEQ}}
}
HighSchoolBench was introduced in the following master’s thesis:
@mastersthesis{rani2026highschoolbench_thesis,
title = {Toward Robust LLM Benchmarking in High School Science Education:
Constructing a QA Corpus from Standardized Tests},
author = {Rani, Krishna},
school = {Gottfried Wilhelm Leibniz Universit{\"a}t Hannover},
year = {2026},
doi = {10.15488/22429},
url = {https://doi.org/10.15488/22429}
}