Claude 4 Opus vs.Qwen 3.5 35B (A3B)
Passed:
Claude 4 Opus 62.7% (52/83)
Qwen 3.5 35B (A3B) 65.9% (54/82)
Average request time:
Claude 4 Opus 17.98s
Qwen 3.5 35B (A3B) 1.28s
Passed:
Claude 4 Opus 62.7% (52/83)
Qwen 3.5 35B (A3B) 65.9% (54/82)
Average request time:
Claude 4 Opus 17.98s
Qwen 3.5 35B (A3B) 1.28s
Summary
Annotation Understanding
Claude 4 Opus
1 / 2
Qwen 3.5 35B (A3B)
1 / 2
CAPTCHA
Claude 4 Opus
2 / 2
Qwen 3.5 35B (A3B)
2 / 2
Color Identification
Claude 4 Opus
1 / 1
Qwen 3.5 35B (A3B)
0 / 1
Defect Detection
Claude 4 Opus
10 / 15
Qwen 3.5 35B (A3B)
12 / 15
Document Understanding
Claude 4 Opus
9 / 10
Qwen 3.5 35B (A3B)
8 / 10
Localization
Claude 4 Opus
0 / 1
Qwen 3.5 35B (A3B)
0 / 1
OCR
Claude 4 Opus
8 / 9
Qwen 3.5 35B (A3B)
8 / 9
Object Counting
Claude 4 Opus
0 / 11
Qwen 3.5 35B (A3B)
2 / 11
Object Detection
Claude 4 Opus
2 / 2
Qwen 3.5 35B (A3B)
0 / 2
Object Measurement
Claude 4 Opus
1 / 1
Qwen 3.5 35B (A3B)
0 / 1
Object Understanding
Claude 4 Opus
7 / 11
Qwen 3.5 35B (A3B)
7 / 11
Receipt Reading
Claude 4 Opus
1 / 1
Qwen 3.5 35B (A3B)
1 / 1
Sign Understanding
Claude 4 Opus
1 / 2
Qwen 3.5 35B (A3B)
2 / 2
Spatial Relations
Claude 4 Opus
9 / 17
Qwen 3.5 35B (A3B)
10 / 17
Web Action Understanding
Claude 4 Opus
4 / 4
Qwen 3.5 35B (A3B)
4 / 4
Contribute a Prompt
Have an idea for a prompt? Open a pull request on the project repository!