GPT-5 (high reasoning) vs.Phi 4 Multimodal
Passed:
GPT-5 (high reasoning) 73.5% (61/83)
Phi 4 Multimodal 49.4% (41/83)
Average request time:
GPT-5 (high reasoning) 26.96s
Phi 4 Multimodal 15.92s
Passed:
GPT-5 (high reasoning) 73.5% (61/83)
Phi 4 Multimodal 49.4% (41/83)
Average request time:
GPT-5 (high reasoning) 26.96s
Phi 4 Multimodal 15.92s
Summary
Annotation Understanding
GPT-5 (high reasoning)
1 / 2
Phi 4 Multimodal
1 / 2
CAPTCHA
GPT-5 (high reasoning)
0 / 2
Phi 4 Multimodal
0 / 2
Color Identification
GPT-5 (high reasoning)
1 / 1
Phi 4 Multimodal
0 / 1
Counting nodes
GPT-5 (high reasoning)
0 / 1
Phi 4 Multimodal
0 / 1
Defect Detection
GPT-5 (high reasoning)
12 / 15
Phi 4 Multimodal
11 / 15
Document Understanding
GPT-5 (high reasoning)
8 / 10
Phi 4 Multimodal
7 / 10
Localization
GPT-5 (high reasoning)
1 / 1
Phi 4 Multimodal
0 / 1
OCR
GPT-5 (high reasoning)
9 / 9
Phi 4 Multimodal
5 / 9
Object Counting
GPT-5 (high reasoning)
3 / 10
Phi 4 Multimodal
0 / 10
Object Detection
GPT-5 (high reasoning)
1 / 2
Phi 4 Multimodal
2 / 2
Object Measurement
GPT-5 (high reasoning)
0 / 2
Phi 4 Multimodal
0 / 2
Object Understanding
GPT-5 (high reasoning)
9 / 11
Phi 4 Multimodal
5 / 11
Receipt Reading
GPT-5 (high reasoning)
1 / 1
Phi 4 Multimodal
1 / 1
Sign Understanding
GPT-5 (high reasoning)
2 / 2
Phi 4 Multimodal
2 / 2
Spatial Relations
GPT-5 (high reasoning)
15 / 17
Phi 4 Multimodal
7 / 17
Web Action Understanding
GPT-5 (high reasoning)
4 / 4
Phi 4 Multimodal
2 / 4
Contribute a Prompt
Have an idea for a prompt? Open a pull request on the project repository!