adjustable glasses illustration

eye chart illustration

clipboard illustration

lenses illustration

GPT-5 vs.Qwen 2.5 VL 7B

Passed:

GPT-5 75.9% (63/83)

Qwen 2.5 VL 7B 55.6% (45/81)

Average request time:

GPT-5 25.28s

Qwen 2.5 VL 7B 45.63s

Summary

Annotation Understanding

GPT-5 1 / 2

Qwen 2.5 VL 7B 1 / 2

CAPTCHA

GPT-5 0 / 2

Qwen 2.5 VL 7B 2 / 2

Color Identification

GPT-5 1 / 1

Qwen 2.5 VL 7B 0 / 1

Counting nodes

GPT-5 0 / 1

Qwen 2.5 VL 7B 0 / 0

Defect Detection

GPT-5 13 / 15

Qwen 2.5 VL 7B 9 / 15

Document Understanding

GPT-5 8 / 10

Qwen 2.5 VL 7B 8 / 10

Localization

GPT-5 0 / 1

Qwen 2.5 VL 7B 1 / 1

OCR

GPT-5 9 / 9

Qwen 2.5 VL 7B 6 / 9

Object Counting

GPT-5 4 / 10

Qwen 2.5 VL 7B 1 / 9

Object Detection

GPT-5 2 / 2

Qwen 2.5 VL 7B 1 / 2

Object Measurement

GPT-5 0 / 2

Qwen 2.5 VL 7B 0 / 2

Object Understanding

GPT-5 9 / 11

Qwen 2.5 VL 7B 5 / 11

Receipt Reading

GPT-5 1 / 1

Qwen 2.5 VL 7B 0 / 1

Sign Understanding

GPT-5 2 / 2

Qwen 2.5 VL 7B 2 / 2

Spatial Relations

GPT-5 15 / 17

Qwen 2.5 VL 7B 11 / 17

Web Action Understanding

GPT-5 4 / 4

Qwen 2.5 VL 7B 4 / 4

Contribute a Prompt

Lightbulb

Have an idea for a prompt? Open a pull request on the project repository!