Qwen3.8-27B
Qwen3.8-27B scores 59.0 out of 100 overall, which puts it in a tie for rank 10–18 of 36 models on the AI Vision Index. It is strongest at Spotting changes and errors (78) and weakest at Counting (47). It beats the average of all 36 models in all six categories. It was tested on a smaller sample than most models, so its margin of error is wider and its single-test scores are rough.
Overall score59.095% interval 51.8–68.0
Rank10–18of 36 models (tied)
Cost per 1,000 answers$7.75
Seconds per answer27.5 s
Thinking effortdefault
Maker Alibaba
WeightsOpen weights
AccessAPI
Tested
Reading52
Counting47
Space, 3D and motion52
Spotting changes and errors78
Diagrams60
Not being fooled65
The bar is the score; the light band is the 95% interval. The overall score is the average of the six.
Reading
Reading a clock60
Reading a meter40
Eye chart20
Reading a label60
Reading a chart50
Test strips80
Adding up coins40
Traffic light from a car70
Counting
Counting objects80
Counting stacked goods0
Counting on a conveyor60
Space, 3D and motion
Which is nearer100
Which way it faces29
Same or mirror image0
Will it topple40
Match the render47
Which bay can I park in?100
Follow the hidden object45
Spotting changes and errors
Spot the change80
Shelf against plan80
Rendering glitch52
Check the invoice100
Diagrams
Follow the cable40
Follow the pipe60
Missing dimension80
Not being fooled
Absent objects100
Saying "I can't tell"100
Optical illusions40
Real or AI-made20
Scores are corrected for guessing: 0 is no better than a guess, 100 is always right. All models on every test
Neighbours on the leaderboard
- GPT-6 Luna 60.7 Compare
- Qwen3.8 Flash 56.4 Compare
Badge for your site
HTML
<a href="https://aivisionindex.com/models/qwen3-8-27b/"><img src="https://aivisionindex.com/badge/qwen3-8-27b.svg" alt="Qwen3.8-27B: ranked #10–18 of 36 on the AI Vision Index"></a>Markdown
[](https://aivisionindex.com/models/qwen3-8-27b/)Results are CC BY 4.0, so you are free to reuse them if you credit the AI Vision Index. All results (JSON)