DeepSeek R1
Browser Sandbox
Loading AI Engine...
Select Model
DeepSeek-R1-Distill (Qwen 1.5B — 900MB)
Llama-3.2 (1B Params — 600MB)
Loading Engine...
Initializing...
Metrics & Hardware
Device
Detecting...
Speed
— t/s
Hello! I'm DeepSeek-R1 running in your browser via WebGPU. Click
Initialize Model
to load the AI, then start chatting!