ON-DEVICE MODELS

Choose your intelligence.

Reviewed LiteRT-LM models that run privately on your device.

Connecting to the reviewed catalog…

◉gemma4

Gemma 4 - 2.3B

Multimodal local model with optional thinking support.

2.4 GBCPU
↗
◉general

Phi 4 - 3.8B

Compact reasoning model from Microsoft.

2.6 GBCPU
↗
◉gemma4

Gemma 4 - 4.5B

Larger multimodal local model with thinking support.

3.4 GBCPU
↗