Gemma 3 4B vs.Llama 4 Maverick 17B
Passed:
Gemma 3 4B 42.2% (35/83)
Llama 4 Maverick 17B 59.0% (49/83)
Average request time:
Gemma 3 4B 15.86s
Llama 4 Maverick 17B 16.22s
Passed:
Gemma 3 4B 42.2% (35/83)
Llama 4 Maverick 17B 59.0% (49/83)
Average request time:
Gemma 3 4B 15.86s
Llama 4 Maverick 17B 16.22s
Summary
Annotation Understanding
Gemma 3 4B
0 / 2
Llama 4 Maverick 17B
0 / 2
CAPTCHA
Gemma 3 4B
2 / 2
Llama 4 Maverick 17B
2 / 2
Color Identification
Gemma 3 4B
0 / 1
Llama 4 Maverick 17B
0 / 1
Defect Detection
Gemma 3 4B
9 / 15
Llama 4 Maverick 17B
11 / 15
Document Understanding
Gemma 3 4B
6 / 10
Llama 4 Maverick 17B
9 / 10
Localization
Gemma 3 4B
0 / 1
Llama 4 Maverick 17B
0 / 1
OCR
Gemma 3 4B
6 / 9
Llama 4 Maverick 17B
5 / 9
Object Counting
Gemma 3 4B
0 / 11
Llama 4 Maverick 17B
3 / 11
Object Detection
Gemma 3 4B
2 / 2
Llama 4 Maverick 17B
2 / 2
Object Measurement
Gemma 3 4B
0 / 1
Llama 4 Maverick 17B
0 / 1
Object Understanding
Gemma 3 4B
5 / 11
Llama 4 Maverick 17B
5 / 11
Receipt Reading
Gemma 3 4B
0 / 1
Llama 4 Maverick 17B
1 / 1
Sign Understanding
Gemma 3 4B
2 / 2
Llama 4 Maverick 17B
2 / 2
Spatial Relations
Gemma 3 4B
4 / 17
Llama 4 Maverick 17B
10 / 17
Web Action Understanding
Gemma 3 4B
2 / 4
Llama 4 Maverick 17B
4 / 4
Contribute a Prompt
Have an idea for a prompt? Open a pull request on the project repository!