Quantization format
UD_Q4_K_S
UD_Q4_K_S has a nominal rate of — bits per weight. We have no measured files carrying this label yet.
From the file· 28 files measured
Nominal bpw
—
from the block layout
Measured average
5.285
28 files
Range
4.521–12.367
varies by architecture
File sizes
4.67 GiB+
up to 548.12 GiB
Real files
one per model, most downloaded first
| Model | Size● | Effective bpw● | vs nominal● |
|---|---|---|---|
| Qwen3.6-35B-A3B | 19.46 GiB | 4.649 | — |
| gemma-4-26B-A4B-it | 15.36 GiB | 4.969 | — |
| Qwen-AgentWorld-35B-A3B | 19.46 GiB | 4.822 | — |
| LTX-2.3 | 13.12 GiB | 12.257 | — |
| GLM-5.2 | 406.46 GiB | 4.635 | — |
| Ornith-1.0-35B | 19.46 GiB | 4.822 | — |
| Qwen3.5-122B-A10B | 68.39 GiB | 4.696 | — |
| Qwen3-Coder-Next | 42.92 GiB | 4.627 | — |
| LTX-2 | 10.83 GiB | 4.929 | — |
| Laguna-S-2.1 | 63.88 GiB | 4.668 | — |
| Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16 | 21.47 GiB | 5.585 | — |
| Step-3.7-Flash | 106.32 GiB | 4.536 | — |
| LFM2.5-8B-A1B | 4.67 GiB | 4.737 | — |
| Ornith-1.0-397B | 213.94 GiB | 4.631 | — |
| Qwen3.5-397B-A17B | 212.33 GiB | 4.521 | — |
| North-Mini-Code-1.0 | 16.81 GiB | 4.736 | — |
| MiniMax-M2.7 | 121.97 GiB | 4.581 | — |
| MiniMax-M3 | 230.71 GiB | 4.641 | — |
| NVIDIA-Nemotron-3-Super-120B-A12B-BF16 | 73.59 GiB | 5.114 | — |
| Mistral-Small-4-119B-2603 | 64.70 GiB | 4.654 | — |
| Huihui-gemma-4-26B-A4B-it-abliterated | 15.27 GiB | 4.940 | — |
| MiMo-V2.5 | 166.56 GiB | 4.604 | — |
| GLM-5.1 | 404.10 GiB | 4.605 | — |
| NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 | 305.18 GiB | 4.677 | — |
| MiMo-V2.5-Pro | 548.12 GiB | 4.601 | — |
●From the filewhat these mean