Claude 4 Sonnet vs.Gemma 3 4B
Passed:
Claude 4 Sonnet 71.1% (59/83)
Gemma 3 4B 42.2% (35/83)
Average request time:
Claude 4 Sonnet 19.72s
Gemma 3 4B 15.86s
Passed:
Claude 4 Sonnet 71.1% (59/83)
Gemma 3 4B 42.2% (35/83)
Average request time:
Claude 4 Sonnet 19.72s
Gemma 3 4B 15.86s
Summary
Annotation Understanding
Claude 4 Sonnet
2 / 2
Gemma 3 4B
0 / 2
CAPTCHA
Claude 4 Sonnet
2 / 2
Gemma 3 4B
2 / 2
Color Identification
Claude 4 Sonnet
0 / 1
Gemma 3 4B
0 / 1
Counting nodes
Claude 4 Sonnet
0 / 1
Gemma 3 4B
0 / 1
Defect Detection
Claude 4 Sonnet
12 / 15
Gemma 3 4B
9 / 15
Document Understanding
Claude 4 Sonnet
9 / 10
Gemma 3 4B
6 / 10
Localization
Claude 4 Sonnet
0 / 1
Gemma 3 4B
0 / 1
OCR
Claude 4 Sonnet
7 / 9
Gemma 3 4B
6 / 9
Object Counting
Claude 4 Sonnet
2 / 10
Gemma 3 4B
0 / 10
Object Detection
Claude 4 Sonnet
2 / 2
Gemma 3 4B
2 / 2
Object Measurement
Claude 4 Sonnet
1 / 2
Gemma 3 4B
0 / 2
Object Understanding
Claude 4 Sonnet
8 / 11
Gemma 3 4B
5 / 11
Receipt Reading
Claude 4 Sonnet
1 / 1
Gemma 3 4B
0 / 1
Sign Understanding
Claude 4 Sonnet
2 / 2
Gemma 3 4B
2 / 2
Spatial Relations
Claude 4 Sonnet
10 / 17
Gemma 3 4B
4 / 17
Web Action Understanding
Claude 4 Sonnet
4 / 4
Gemma 3 4B
2 / 4
Contribute a Prompt
Have an idea for a prompt? Open a pull request on the project repository!