MLX port of flux-schnell is surprisingly fast. With q4 weights and CFG-free guidance: - 1024×1024, 4 steps: 5.2s on M3 Max (64G) - 512×512, 4 steps: 1.4s - 768×768 batched (n=2), 4 steps: 6.9s → 3.45s/img That's faster …