Phi 4 Multimodal vs.OpenAI o3-pro
Passed:
Phi 4 Multimodal 56.1% (46/82)
OpenAI o3-pro 76.8% (63/82)
Average request time:
Phi 4 Multimodal 4.08s
OpenAI o3-pro 39.50s
Passed:
Phi 4 Multimodal 56.1% (46/82)
OpenAI o3-pro 76.8% (63/82)
Average request time:
Phi 4 Multimodal 4.08s
OpenAI o3-pro 39.50s
Summary
Annotation Understanding
Phi 4 Multimodal
2 / 2
OpenAI o3-pro
2 / 2
CAPTCHA
Phi 4 Multimodal
0 / 2
OpenAI o3-pro
0 / 2
Color Identification
Phi 4 Multimodal
1 / 1
OpenAI o3-pro
1 / 1
Counting nodes
Phi 4 Multimodal
0 / 1
OpenAI o3-pro
0 / 1
Defect Detection
Phi 4 Multimodal
12 / 15
OpenAI o3-pro
12 / 15
Document Understanding
Phi 4 Multimodal
4 / 10
OpenAI o3-pro
9 / 10
Localization
Phi 4 Multimodal
0 / 1
OpenAI o3-pro
1 / 1
OCR
Phi 4 Multimodal
6 / 9
OpenAI o3-pro
9 / 9
Object Counting
Phi 4 Multimodal
0 / 10
OpenAI o3-pro
3 / 10
Object Detection
Phi 4 Multimodal
1 / 2
OpenAI o3-pro
2 / 2
Object Measurement
Phi 4 Multimodal
0 / 2
OpenAI o3-pro
0 / 2
Object Understanding
Phi 4 Multimodal
8 / 10
OpenAI o3-pro
8 / 10
Receipt Reading
Phi 4 Multimodal
1 / 1
OpenAI o3-pro
1 / 1
Sign Understanding
Phi 4 Multimodal
1 / 2
OpenAI o3-pro
2 / 2
Spatial Relations
Phi 4 Multimodal
9 / 17
OpenAI o3-pro
15 / 17
Web Action Understanding
Phi 4 Multimodal
3 / 4
OpenAI o3-pro
4 / 4
Contribute a Prompt
Have an idea for a prompt? Open a pull request on the project repository!