GLM 4.6v vs.SmolVLM2 2.2B
Passed:
GLM 4.6v 68.0% (51/75)
SmolVLM2 2.2B 36.6% (30/82)
Average request time:
GLM 4.6v 12.37s
SmolVLM2 2.2B 4.16s
Passed:
GLM 4.6v 68.0% (51/75)
SmolVLM2 2.2B 36.6% (30/82)
Average request time:
GLM 4.6v 12.37s
SmolVLM2 2.2B 4.16s
Summary
Annotation Understanding
GLM 4.6v
0 / 0
SmolVLM2 2.2B
2 / 2
CAPTCHA
GLM 4.6v
2 / 2
SmolVLM2 2.2B
0 / 2
Color Identification
GLM 4.6v
0 / 1
SmolVLM2 2.2B
0 / 1
Defect Detection
GLM 4.6v
11 / 15
SmolVLM2 2.2B
9 / 15
Document Understanding
GLM 4.6v
7 / 8
SmolVLM2 2.2B
3 / 10
Localization
GLM 4.6v
1 / 1
SmolVLM2 2.2B
0 / 1
OCR
GLM 4.6v
7 / 9
SmolVLM2 2.2B
2 / 9
Object Counting
GLM 4.6v
2 / 9
SmolVLM2 2.2B
1 / 11
Object Detection
GLM 4.6v
0 / 2
SmolVLM2 2.2B
1 / 2
Object Measurement
GLM 4.6v
0 / 1
SmolVLM2 2.2B
0 / 1
Object Understanding
GLM 4.6v
8 / 10
SmolVLM2 2.2B
3 / 11
Receipt Reading
GLM 4.6v
1 / 1
SmolVLM2 2.2B
0 / 1
Sign Understanding
GLM 4.6v
2 / 2
SmolVLM2 2.2B
2 / 2
Spatial Relations
GLM 4.6v
9 / 17
SmolVLM2 2.2B
10 / 17
Web Action Understanding
GLM 4.6v
4 / 4
SmolVLM2 2.2B
0 / 4
Contribute a Prompt
Have an idea for a prompt? Open a pull request on the project repository!