Live · On-device in-browser inference

On-Device AI Benchmarks
for the Open Web

LiteRT.js, ONNX Runtime Web, WebNN, WebGPU, and Wasm. Run real ML models in your browser and compare performance across backends - no install, no server.

WebNN WebGPU Wasm .tflite .litertlm .task .onnx LiteRT.js ONNX Runtime Web HuggingFace
Latest Results

Recent Model Updates

Loading models...

Cart

No models selected.

to save models as a recipe.