Show HN: We compressed a 31B video+audio diffusion model to 14B, 6 steps, INT8

Automated excerpt

Speed (5.2 s clip, 6 steps): 8.7 s on 4× B200, 36 s on 1× H200, 92 s on 1× RTX 5090. It runs on 24 GB and 32 GB GPUs in PCs with 16 GB of system RAM. 16 GB GPUs are not measured yet. Download the 15.7 GB qwen3vl_32b_minimax_h3_nvfp4_awq text encoder.

Selected automatically from source text; not independently written or fact-checked. Read the original for full context.