Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

ScienceQA

Benchmarks

Task NameDataset NameSOTA ResultTrend
Science Question AnsweringScienceQA
Accuracy98.5
502
Visual Question AnsweringScienceQA
Accuracy98.18
370
Science Question AnsweringScienceQA IMG
Accuracy98.2
294
Science Question AnsweringScienceQA (SQA)
Accuracy88.8
273
Science Question AnsweringScienceQA (test)
Average Accuracy98.3
245
Scientific Question AnsweringScienceQA image
Accuracy98
184
Science Question AnsweringScienceQA SQA-IMG
Accuracy92.8
139
Multimodal Science Question AnsweringScienceQA IMG
Accuracy95.78
131
Visual Question AnsweringScienceQA (test)
Accuracy98.2
113
Science Question AnsweringScienceQA SQA-I
Accuracy96.6
103
Question AnsweringScienceQA
Accuracy96
77
Science Question AnsweringScienceQA IMG (test)
Accuracy98.4
74
Multimodal Question AnsweringScienceQA (test)
Accuracy92.53
65
Science Question AnsweringScienceQA
IMG Score88.6
64
Scientific Question AnsweringScienceQA
Accuracy88.6
61
Visual Question AnsweringScienceQA img
Accuracy86.2
52
Science Question AnsweringScienceQA Image
Score74.2
51
Visual Question AnsweringScienceQA image
Accuracy75.1
51
Visual Question AnsweringScienceQA Image (test)
Accuracy96.2
45
Visual Question AnsweringScienceQA (SQA)
SQA Accuracy89.3
43
Multimodal Question AnsweringScienceQA
Accuracy97.8
41
Out-of-scope refusalScienceQA out-of-scope (test)
Refusal Rate0
40
Multimodal Science Question AnsweringScienceQA
Overall Average Score88.4
36
Multi-step ReasoningScienceQA
Accuracy92.72
35
Visual Question AnsweringScienceQA (SQAI)
Accuracy74.2
35
Showing 25 of 79 rows