- (01)
27–35B local models on MLX
- (02)
Captures the screen and describes it with a local multimodal model in ~4 s
- (03)
Qwen3-Omni 30B running locally with audio input at ~76–100 tok/s
27–35B local models on MLX
Captures the screen and describes it with a local multimodal model in ~4 s
Qwen3-Omni 30B running locally with audio input at ~76–100 tok/s