Mistral Medium 3 vs.Mistral Small 3.1 24B
Passed:
Mistral Medium 3 68.7% (57/83)
Mistral Small 3.1 24B 44.6% (37/83)
Average request time:
Mistral Medium 3 13.13s
Mistral Small 3.1 24B 15.16s
Passed:
Mistral Medium 3 68.7% (57/83)
Mistral Small 3.1 24B 44.6% (37/83)
Average request time:
Mistral Medium 3 13.13s
Mistral Small 3.1 24B 15.16s
Summary
Annotation Understanding
Mistral Medium 3
1 / 2
Mistral Small 3.1 24B
0 / 2
CAPTCHA
Mistral Medium 3
2 / 2
Mistral Small 3.1 24B
2 / 2
Color Identification
Mistral Medium 3
0 / 1
Mistral Small 3.1 24B
1 / 1
Counting nodes
Mistral Medium 3
0 / 1
Mistral Small 3.1 24B
0 / 1
Defect Detection
Mistral Medium 3
12 / 15
Mistral Small 3.1 24B
8 / 15
Document Understanding
Mistral Medium 3
8 / 10
Mistral Small 3.1 24B
5 / 10
Localization
Mistral Medium 3
1 / 1
Mistral Small 3.1 24B
0 / 1
OCR
Mistral Medium 3
9 / 9
Mistral Small 3.1 24B
7 / 9
Object Counting
Mistral Medium 3
2 / 10
Mistral Small 3.1 24B
0 / 10
Object Detection
Mistral Medium 3
2 / 2
Mistral Small 3.1 24B
1 / 2
Object Measurement
Mistral Medium 3
0 / 2
Mistral Small 3.1 24B
0 / 2
Object Understanding
Mistral Medium 3
9 / 11
Mistral Small 3.1 24B
4 / 11
Receipt Reading
Mistral Medium 3
1 / 1
Mistral Small 3.1 24B
1 / 1
Sign Understanding
Mistral Medium 3
2 / 2
Mistral Small 3.1 24B
0 / 2
Spatial Relations
Mistral Medium 3
7 / 17
Mistral Small 3.1 24B
6 / 17
Web Action Understanding
Mistral Medium 3
4 / 4
Mistral Small 3.1 24B
4 / 4
Contribute a Prompt
Have an idea for a prompt? Open a pull request on the project repository!