DeepSeek V4.1 Flash
DeepSeek V4.1 Flash scores 44.4 out of 100 overall, which puts it in a tie for rank 19–29 of 36 models on the AI Vision Index. It is strongest at Spotting changes and errors (61) and weakest at Counting (17). It beats the average of all 36 models in one of the six categories.
Overall score44.495% interval 43.2–45.8
Rank19–29of 36 models (tied)
Cost per 1,000 answers$1.43
Seconds per answer10.9 s
Thinking effortdefault
Maker DeepSeek
WeightsOpen weights
AccessAPI
Tested
Reading41
Counting17
Space, 3D and motion42
Spotting changes and errors61
Diagrams54
Not being fooled52
The bar is the score; the light band is the 95% interval. The overall score is the average of the six.
Reading
Reading a clock26
Reading a meter36
Eye chart20
Reading a label68
Reading a chart50
Test strips45
Adding up coins32
Traffic light from a car48
Counting
Counting objects29
Counting stacked goods2
Counting on a conveyor22
Space, 3D and motion
Which is nearer65
Which way it faces36
Same or mirror image0
Will it topple24
Match the render42
Which bay can I park in?89
Follow the hidden object37
Spotting changes and errors
Spot the change62
Shelf against plan58
Rendering glitch30
Check the invoice92
Diagrams
Follow the cable52
Follow the pipe32
Missing dimension78
Not being fooled
Absent objects86
Saying "I can't tell"90
Optical illusions0
Real or AI-made31
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
- GLM-5V-Turbo 46.8 Compare
- GLM-5.3 Flash 43.2 Compare
Badge for your site
HTML
<a href="https://aivisionindex.com/models/deepseek-v4-1-flash/"><img src="https://aivisionindex.com/badge/deepseek-v4-1-flash.svg" alt="DeepSeek V4.1 Flash: ranked #19–29 of 36 on the AI Vision Index"></a>Markdown
[](https://aivisionindex.com/models/deepseek-v4-1-flash/)Results are CC BY 4.0, so you are free to reuse them if you credit the AI Vision Index. All results (JSON)