8–12GB | RTX 50 | all 208 matrices NVFP4

16–24GB | RTX 50 | 170×NVFP4 + 30×BF16 + BF16 refiner

8GB + CPU offload | RTX 30/40 | all 208 matrices W4

12GB | RTX 30/40 | 200×W4 + 8×INT8 refiner

16GB | RTX 30/40 | 86×W8 + 114×W4 + BF16 refiner

24GB | RTX 30/40 | 185×INT8 ConvRot + 23×BF16

24GB | RTX 30/40 | 170×INT8 ConvRot + 38×BF16

32GB+ | RTX 30/40 | 145×INT8 ConvRot + 63×BF16

Remove superseded long checkpoint names

Update FL2VA/SHA256SUMS for short checkpoint names

Add short canonical checkpoint names

Add minimax-h3-fl2va-table-k16-g4097-nvfp4-compact-blackwell.safetensors

Add minimax-h3-fl2va-table-k16-g4097-nvfp4-quality21-blackwell.safetensors

Add minimax-h3-fl2va-table-k16-g4097-w4-offload8.safetensors

Add minimax-h3-fl2va-table-k16-g4097-w4-compact12.safetensors

Add minimax-h3-fl2va-table-k16-g4097-w8w4-balanced16.safetensors

Add minimax-h3-fl2va-int8-lean-convrot-table-k16-g4097-quality20.safetensors

Add minimax-h3-fl2va-int8-lean-convrot-table-k16-g4097-quality22.safetensors

Move dynamic-time separate-QKV checkpoints to dedicated repository

Add files using upload-large-folder tool

Make stock-compatible table-k16 checkpoint the default release

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add model card, ComfyUI patch, validation reports, and tools