Datasets:

mipbench-anon
/

mipbench

id large_stringlengths 5 8	question_type large_stringclasses 3 values	question large_stringlengths 40 740	choices large_stringlengths 23 4.31k ⌀	answer large_stringclasses 370 values	source large_stringclasses 9 values	class large_stringclasses 5 values	raw_id large_stringlengths 3 40
com_0	mc	<image> <image> Is the material of the forks in image 1 the same as the material of the forks in image 2?	{"A": "Yes", "B": "No", "C": "Not enough information to determine"}	B	Mantis-Eval	Comparison	cw_11
com_1	mc	<image> <image> These two images are from the dashcam of the same car. Please determine whether the car was on the same road when these two pictures were taken.	{"A": "Definitely not on the same road", "B": "Definitely on the same road", "C": "Possibly on the same road", "D": "Possibly not on the same road"}	B	MMSI-Bench	Comparison	mmsi_586
com_2	mc	<image> <image> Are there any patterns on the bag in image 1 that match with the bag in image 2?	{"A": "Yes", "B": "No", "C": "Not enough information to determine"}	B	Mantis-Eval	Comparison	cw_13
com_3	mc	<image> <image> Are there any patterns on the bag in image 1 that match with the bag in image 2?	{"A": "Yes", "B": "No", "C": "Not enough information to determine"}	A	Mantis-Eval	Comparison	cw_14
com_4	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	yes	MIRB	Comparison	attribute_0
com_5	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	no	MIRB	Comparison	attribute_1
com_6	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	yes	MIRB	Comparison	attribute_2
com_7	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	yes	MIRB	Comparison	attribute_3
com_8	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	no	MIRB	Comparison	attribute_4
com_9	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	no	MIRB	Comparison	attribute_5
com_10	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	no	MIRB	Comparison	attribute_6
com_11	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	no	MIRB	Comparison	attribute_7
com_12	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	no	MIRB	Comparison	attribute_8
com_13	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	yes	MIRB	Comparison	attribute_9
com_14	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	no	MIRB	Comparison	attribute_10
com_15	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	yes	MIRB	Comparison	attribute_11
com_16	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	yes	MIRB	Comparison	attribute_12
com_17	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	yes	MIRB	Comparison	attribute_13
com_18	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	no	MIRB	Comparison	attribute_14
com_19	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	yes	MIRB	Comparison	attribute_15
com_20	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	no	MIRB	Comparison	attribute_17
com_21	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	no	MIRB	Comparison	attribute_18
com_22	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	yes	MIRB	Comparison	attribute_19
com_23	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	yes	MIRB	Comparison	attribute_20
com_24	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	no	MIRB	Comparison	attribute_21
com_25	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	yes	MIRB	Comparison	attribute_22
com_26	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	yes	MIRB	Comparison	attribute_23
com_27	yes_no	<image> <image> <image> <image> Do all the images contain the same kind of object?	null	no	MIRB	Comparison	attribute_24
com_28	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	no	MIRB	Comparison	attribute_50
com_29	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	yes	MIRB	Comparison	attribute_51
com_30	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	yes	MIRB	Comparison	attribute_52
com_31	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	no	MIRB	Comparison	attribute_53
com_32	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	yes	MIRB	Comparison	attribute_55
com_33	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	no	MIRB	Comparison	attribute_56
com_34	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	no	MIRB	Comparison	attribute_57
com_35	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	no	MIRB	Comparison	attribute_58
com_36	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	no	MIRB	Comparison	attribute_59
com_37	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	no	MIRB	Comparison	attribute_60
com_38	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	no	MIRB	Comparison	attribute_61
com_39	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	yes	MIRB	Comparison	attribute_62
com_40	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	yes	MIRB	Comparison	attribute_63
com_41	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	yes	MIRB	Comparison	attribute_64
com_42	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	yes	MIRB	Comparison	attribute_65
com_43	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	no	MIRB	Comparison	attribute_66
com_44	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	yes	MIRB	Comparison	attribute_67
com_45	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	no	MIRB	Comparison	attribute_68
com_46	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	yes	MIRB	Comparison	attribute_69
com_47	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	no	MIRB	Comparison	attribute_70
com_48	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	yes	MIRB	Comparison	attribute_71
com_49	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	yes	MIRB	Comparison	attribute_72
com_50	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	yes	MIRB	Comparison	attribute_73
com_51	yes_no	<image> <image> <image> <image> Do all the images belong to the same visual domain?	null	yes	MIRB	Comparison	attribute_74
com_52	mc	<image> <image> <image> Do these images show 3 different geographical places?	{"A": "Yes, they show different 3 geographical places", "B": "No, two of them show the same geographical places", "C": "No, all of them show the same geographical places"}	A	Mantis-Eval	Comparison	xh_11
com_53	yes_no	<image> <image> Do you expect both images to show up in the same section of a shopping center? (answer with yes or no)	null	no	Mantis-Eval	Comparison	w_30
com_54	mc	<image> <image> Do you expect both images to show up in the same section of a shopping center?	{"A": "Yes, they both belong to men's clothing section", "B": "Yes, they both belong to women's clothing section", "C": "No, they belong to different sections"}	B	Mantis-Eval	Comparison	w_31
com_55	mc	<image> <image> Do we have the same number and same type of item in both images?	{"A": "same number and same type of item", "B": "same number but different type of item", "C": "different number but same type of item", "D": "different number and different type of item"}	A	Mantis-Eval	Comparison	w_38
com_56	mc	<image> <image> What are the differences between the two images?	{"A": "The trees in the background have more leaves, and there are fewer people in the park.", "B": "There are no longer people standing on the sidewalk, and the car has moved from the parking lot.", "C": "None of the choices provided", "D": "The building in the background has changed color, and there is a new bench in...	B	MUIRBENCH	Comparison	794
com_57	mc	<image> <image> What are the differences between the two images?	{"A": "None of the choices provided", "B": "The building in the background has changed color, and there is a new bench in the park.", "C": "The trees in the background have more leaves, and there are fewer people in the park.", "D": "The car in the parking lot has changed color, and there are more people on the sidewal...	A	MUIRBENCH	Comparison	795
com_58	mc	<image> <image> What similarities can you spot in the behaviors of the animals in and ?	{"A": "both of them are not happy", "B": "both of them looks excited", "C": "both of them looks tired", "D": "both of them looks angry"}	D	Mantis-Eval	Comparison	xh_48
com_59	yes_no	<image> <image> Do the images represent different seasons?	null	yes	Mantis-Eval	Comparison	m_9
com_60	mc	<image> <image> Do the images represent different seasons?	{"A": "Yes", "B": "No", "C": "Not enough information to determine"}	A	Mantis-Eval	Comparison	cw_20
com_61	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "the picture on the right has more people", "B": "The composition of the image is dominated by bright colors and geometric shapes, creating a visual contrast between order and chaos."}	A	MMIU-Benchmark	Comparison	mmiu_3947
com_62	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "the guy is no longer by his car he is further towards the middle of the parking lot", "B": "The man is now riding a bicycle in the park."}	A	MMIU-Benchmark	Comparison	mmiu_3834
com_63	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "the 2 people on the sidewalk are gone", "B": "A colorful mural of a cityscape fills the entire wall, with no people in sight."}	A	MMIU-Benchmark	Comparison	mmiu_3835
com_64	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "Empty streets with no signs of activity", "B": "less cars parked"}	B	MMIU-Benchmark	Comparison	mmiu_3836
com_65	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "A group of friends enjoying a picnic in the park", "B": "the silver car left the parking lot"}	B	MMIU-Benchmark	Comparison	mmiu_3837
com_66	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "two pepole standing by building is no longer there", "B": "A colorful hot air balloon floating in the sky over a peaceful lake with a small cabin on the shore."}	A	MMIU-Benchmark	Comparison	mmiu_3838
com_67	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "there are now two people by the yellow poles", "B": "A colorful parrot sits on a tree branch, preening its feathers."}	A	MMIU-Benchmark	Comparison	mmiu_3840
com_68	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "there s one more car in the group of cars on the far left of the parking lot and also in the group in the center of the lot", "B": "A flock of birds is flying in the clear blue sky above the parking lot"}	A	MMIU-Benchmark	Comparison	mmiu_3841
com_69	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "The color of the sky is a vibrant shade of purple, and the ground is covered in a shimmering layer of silver dust.", "B": "there are 6 people on the right side and none on the left"}	B	MMIU-Benchmark	Comparison	mmiu_3843
com_70	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "Both images show a beautiful landscape with colorful flowers and lush greenery. In the distance, a majestic mountain range rises against the horizon, creating a stunning backdrop for the scene.", "B": "there are more people in the one on the right than on the left"}	B	MMIU-Benchmark	Comparison	mmiu_3845
com_71	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "The road is now empty with no signs of any vehicles or people.", "B": "a person on a motor cycle is no longer in the image"}	B	MMIU-Benchmark	Comparison	mmiu_3847
com_72	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "four people have appeared on the picture", "B": "The image shows a colorful landscape with a beautiful sunset."}	A	MMIU-Benchmark	Comparison	mmiu_3848
com_73	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "The colorful mural on the wall seems to come to life as the vibrant hues blend together in a mesmerizing dance.", "B": "the people on the stairs are closer"}	B	MMIU-Benchmark	Comparison	mmiu_3850
com_74	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "more pedestrians visiable", "B": "A group of colorful hot air balloons soaring through the sky"}	A	MMIU-Benchmark	Comparison	mmiu_3851
com_75	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "there was six people standing around and now there are two", "B": "The room was brightly lit with colorful decorations and lively music playing in the background."}	A	MMIU-Benchmark	Comparison	mmiu_3852
com_76	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "no noticeable differences", "B": "A red apple sits on a wooden table, next to a stack of books."}	A	MMIU-Benchmark	Comparison	mmiu_3853
com_77	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "A colorful meadow with a blue sky and fluffy clouds.", "B": "there is no difference"}	B	MMIU-Benchmark	Comparison	mmiu_3854
com_78	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "no noticeable changes have occurred", "B": "An interesting juxtaposition of colors and shadows is evident."}	A	MMIU-Benchmark	Comparison	mmiu_3855
com_79	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "the white car is no longer on the road adjacent to the car parking lot", "B": "A group of people are playing on a sunny beach near the ocean."}	A	MMIU-Benchmark	Comparison	mmiu_3856
com_80	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "A cat is sitting on a tree branch", "B": "there are two people standing in the street"}	B	MMIU-Benchmark	Comparison	mmiu_3857
com_81	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "The sky is filled with colorful hot air balloons.", "B": "the man is road"}	B	MMIU-Benchmark	Comparison	mmiu_3860
com_82	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "more cars in picture", "B": "A group of people playing frisbee in a park."}	A	MMIU-Benchmark	Comparison	mmiu_3861
com_83	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "all the people are new", "B": "The room is decorated with vintage furniture and colorful murals."}	A	MMIU-Benchmark	Comparison	mmiu_3863
com_84	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "there are less pedestrians", "B": "The buildings in the background have a unique architecture that stands out in the cityscape."}	A	MMIU-Benchmark	Comparison	mmiu_3865
com_85	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "The blue car is now the only one in the parking lot", "B": "the red car is no longer there"}	B	MMIU-Benchmark	Comparison	mmiu_3866
com_86	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "the group of people has moved to the left", "B": "A colorful mural adorns the side of the building, depicting various abstract shapes and patterns."}	A	MMIU-Benchmark	Comparison	mmiu_3867
com_87	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "the white truck is not there anumore", "B": "A group of birds is flying in the clear blue sky."}	A	MMIU-Benchmark	Comparison	mmiu_3868
com_88	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "The clouds overhead resemble an elaborate maze, with no clear path to navigate.", "B": "the people in the lot have only moved slightly"}	B	MMIU-Benchmark	Comparison	mmiu_3869
com_89	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "A group of people are enjoying a picnic in a park", "B": "the van has backed out the parking space it was in"}	B	MMIU-Benchmark	Comparison	mmiu_3871
com_90	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "the white car is no longer there", "B": "An unexpected gathering of birds occurred in the vicinity."}	A	MMIU-Benchmark	Comparison	mmiu_3872
com_91	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "The van seems to be covered in graffiti and shows signs of wear and tear, while a squirrel is perched on top of it.", "B": "there appears to be a person standing next to the van which is parked closest to the building"}	B	MMIU-Benchmark	Comparison	mmiu_3873
com_92	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "the group of people have changed places in the circle", "B": "The sun is setting, casting an orange glow across the landscape."}	A	MMIU-Benchmark	Comparison	mmiu_3874
com_93	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "A group of birds is flying over the lake between the trees.", "B": "there are two fewer people standing near the nellow poles"}	B	MMIU-Benchmark	Comparison	mmiu_3875
com_94	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "people are in the photo in after", "B": "The photo features a vibrant sunset over a peaceful lake with a lone boat drifting across the water."}	A	MMIU-Benchmark	Comparison	mmiu_3876
com_95	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "the two people are no longer there near the dividers", "B": "The empty chairs overlook the vast desert landscape."}	A	MMIU-Benchmark	Comparison	mmiu_3877
com_96	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "there are 6 people rather than 2 in the drive", "B": "The picture shows a scenic mountain landscape with a lake, instead of an urban cityscape."}	A	MMIU-Benchmark	Comparison	mmiu_3878
com_97	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "The birds are gathering on the roof to form a plan for world domination.", "B": "a truck in the corner has dissapeared"}	B	MMIU-Benchmark	Comparison	mmiu_3880
com_98	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "A group of friends enjoying a picnic in the park", "B": "only two people outside of parking lot"}	B	MMIU-Benchmark	Comparison	mmiu_3881
com_99	mc	<image> <image> The following is a description of the differences between two pictures. Which one is incorrect?	{"A": "A flock of birds is flying high in the sky, casting shadows on the ground.", "B": "a car is leaving the parking lot near the top of the picture to the left of the white box truck"}	B	MMIU-Benchmark	Comparison	mmiu_3882

End of preview. Expand in Data Studio

MIPBench

MIPBench is an evaluation-only benchmark for measuring position sensitivity in position-invariant multi-image visual question answering. Each example contains multiple images, a question, an answer, and provenance metadata. The benchmark is intended to evaluate whether a vision-language model changes its answer when the input images are permuted while the ground-truth answer remains unchanged.

Dataset Summary

MIPBench contains 6,994 position-invariant multi-image VQA examples derived from nine public source benchmarks. The examples are filtered by a three-stage pipeline and manual review to remove questions whose correct answer depends on image order.

The benchmark is organized into five task categories and three answer formats. It is designed for evaluating accuracy, consistency rate, and flip rate under image permutations.

Intended Use

MIPBench is intended for controlled evaluation of position sensitivity in multi-image vision-language models. It can be used to test whether model predictions remain stable when input images are reordered in tasks where the ground-truth answer should remain unchanged.

Out-of-Scope Use

MIPBench is not intended for model training, general-purpose VLM ranking, or deployment qualification in high-stakes domains such as medicine, transportation, law, or safety-critical decision making.

Data Fields

id: MIPBench example identifier.
question_type: answer format, such as multiple choice, yes/no, or short answer.
question: the multi-image question.
choices: answer options for multiple-choice examples, otherwise null.
answer: ground-truth answer.
images: list of input images.
source: source benchmark.
class: MIPBench taxonomy category.
raw_id: original source-benchmark example identifier.

Provenance

MIPBench is derived from nine public multi-image benchmarks. Each retained example preserves its source benchmark and original identifier through the source and raw_id fields.

Limitations

MIPBench is a high-precision benchmark for position-invariant multi-image VQA, not a comprehensive benchmark of general multi-image reasoning. Because it is derived from existing public benchmarks, it may inherit their linguistic, geographic, demographic, visual-domain, and licensing biases. Questions are English-only. Some residual ambiguity in position invariance may remain despite filtering and manual review.

Downloads last month: 56