OpenAI O4 Mini vs.Claude 3.5 Haiku
Passed:
OpenAI O4 Mini 79.3% (65/82)
Claude 3.5 Haiku 67.1% (55/82)
Average request time:
OpenAI O4 Mini 25.90s
Claude 3.5 Haiku 12.50s
Passed:
OpenAI O4 Mini 79.3% (65/82)
Claude 3.5 Haiku 67.1% (55/82)
Average request time:
OpenAI O4 Mini 25.90s
Claude 3.5 Haiku 12.50s
Summary
Annotation Understanding
OpenAI O4 Mini
2 / 2
Claude 3.5 Haiku
1 / 2
CAPTCHA
OpenAI O4 Mini
0 / 2
Claude 3.5 Haiku
0 / 2
Color Identification
OpenAI O4 Mini
1 / 1
Claude 3.5 Haiku
1 / 1
Counting nodes
OpenAI O4 Mini
0 / 1
Claude 3.5 Haiku
0 / 1
Defect Detection
OpenAI O4 Mini
13 / 15
Claude 3.5 Haiku
10 / 15
Document Understanding
OpenAI O4 Mini
5 / 9
Claude 3.5 Haiku
5 / 9
Find an author
OpenAI O4 Mini
0 / 1
Claude 3.5 Haiku
0 / 1
Localization
OpenAI O4 Mini
1 / 1
Claude 3.5 Haiku
0 / 1
OCR
OpenAI O4 Mini
9 / 9
Claude 3.5 Haiku
7 / 9
Object Counting
OpenAI O4 Mini
5 / 10
Claude 3.5 Haiku
6 / 10
Object Detection
OpenAI O4 Mini
2 / 2
Claude 3.5 Haiku
2 / 2
Object Measurement
OpenAI O4 Mini
1 / 2
Claude 3.5 Haiku
0 / 2
Object Understanding
OpenAI O4 Mini
9 / 10
Claude 3.5 Haiku
8 / 10
Receipt Reading
OpenAI O4 Mini
1 / 1
Claude 3.5 Haiku
1 / 1
Sign Understanding
OpenAI O4 Mini
2 / 2
Claude 3.5 Haiku
2 / 2
Spatial Relations
OpenAI O4 Mini
15 / 17
Claude 3.5 Haiku
13 / 17
Web Action Understanding
OpenAI O4 Mini
4 / 4
Claude 3.5 Haiku
4 / 4
Contribute a Prompt
Have an idea for a prompt? Open a pull request on the project repository!