U Local AI
Benchmark · on-device
Chat
This engine can't run here
Engine under test
Engine
Chrome · Gemini Nano
WebLLM · WebGPU
Open model
Run benchmark
Stop
Include heavy tasks
0 / 15
Result
Copy
0
%
Model load
-
First token
-
Speed
-
tok/s
Context peak
-
tok
Tasks
Your models, ranked
Clear
Model
Score
First token
tok/s
How the score works
Settings
Language
Auto
English
Português
Appearance
System
Light
Dark
About
All settings, in the chat
Translator