Text Generation
MLX
Safetensors
qwen3
conversational
8-bit precision