Claude 4.1 Opus vs.SmolVLM2 2.2B
Passed:
Claude 4.1 Opus 61.4% (51/83)
SmolVLM2 2.2B 36.6% (30/82)
Average request time:
Claude 4.1 Opus 17.48s
SmolVLM2 2.2B 4.16s
Passed:
Claude 4.1 Opus 61.4% (51/83)
SmolVLM2 2.2B 36.6% (30/82)
Average request time:
Claude 4.1 Opus 17.48s
SmolVLM2 2.2B 4.16s
Summary
Annotation Understanding
Claude 4.1 Opus
1 / 2
SmolVLM2 2.2B
2 / 2
CAPTCHA
Claude 4.1 Opus
2 / 2
SmolVLM2 2.2B
0 / 2
Color Identification
Claude 4.1 Opus
1 / 1
SmolVLM2 2.2B
0 / 1
Defect Detection
Claude 4.1 Opus
9 / 15
SmolVLM2 2.2B
9 / 15
Document Understanding
Claude 4.1 Opus
9 / 10
SmolVLM2 2.2B
3 / 10
Localization
Claude 4.1 Opus
0 / 1
SmolVLM2 2.2B
0 / 1
OCR
Claude 4.1 Opus
8 / 9
SmolVLM2 2.2B
2 / 9
Object Counting
Claude 4.1 Opus
0 / 11
SmolVLM2 2.2B
1 / 11
Object Detection
Claude 4.1 Opus
1 / 2
SmolVLM2 2.2B
1 / 2
Object Measurement
Claude 4.1 Opus
0 / 1
SmolVLM2 2.2B
0 / 1
Object Understanding
Claude 4.1 Opus
7 / 11
SmolVLM2 2.2B
3 / 11
Receipt Reading
Claude 4.1 Opus
1 / 1
SmolVLM2 2.2B
0 / 1
Sign Understanding
Claude 4.1 Opus
2 / 2
SmolVLM2 2.2B
2 / 2
Spatial Relations
Claude 4.1 Opus
11 / 17
SmolVLM2 2.2B
10 / 17
Web Action Understanding
Claude 4.1 Opus
4 / 4
SmolVLM2 2.2B
0 / 4
Contribute a Prompt
Have an idea for a prompt? Open a pull request on the project repository!