Quantization format
UD_Q4_K_M
UD_Q4_K_M has a nominal rate of — bits per weight. We have no measured files carrying this label yet.
From the file· 32 files measured
Nominal bpw
—
from the block layout
Measured average
5.947
32 files
Range
4.841–14.358
varies by architecture
File sizes
4.96 GiB+
up to 586.37 GiB
Real files
one per model, most downloaded first
| Model | Size● | Effective bpw● | vs nominal● |
|---|---|---|---|
| Qwen3.6-35B-A3B | 20.61 GiB | 4.925 | — |
| gemma-4-26B-A4B-it | 15.78 GiB | 5.108 | — |
| Qwen-AgentWorld-35B-A3B | 20.61 GiB | 5.109 | — |
| LTX-2.3 | 15.19 GiB | 14.183 | — |
| GLM-5.2 | 433.83 GiB | 4.947 | — |
| Ornith-1.0-35B | 20.61 GiB | 5.109 | — |
| Qwen3.5-122B-A10B | 72.89 GiB | 5.005 | — |
| Qwen3-Coder-Next | 45.92 GiB | 4.950 | — |
| Laguna-S-2.1 | 68.10 GiB | 4.976 | — |
| Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16 | 22.25 GiB | 5.788 | — |
| Step-3.7-Flash | 113.71 GiB | 4.851 | — |
| LFM2.5-8B-A1B | 4.96 GiB | 5.028 | — |
| Ornith-1.0-397B | 228.19 GiB | 4.940 | — |
| Qwen3.5-397B-A17B | 227.33 GiB | 4.841 | — |
| North-Mini-Code-1.0 | 17.88 GiB | 5.040 | — |
| MiniMax-M2.7 | 130.54 GiB | 4.903 | — |
| Nex-N2-mini | 20.60 GiB | 5.039 | — |
| MiniMax-M3 | 245.90 GiB | 4.946 | — |
| NVIDIA-Nemotron-3-Super-120B-A12B-BF16 | 76.87 GiB | 5.342 | — |
| Mistral-Small-4-119B-2603 | 68.70 GiB | 4.942 | — |
| Huihui-gemma-4-26B-A4B-it-abliterated | 15.71 GiB | 5.084 | — |
| MiMo-V2.5 | 177.81 GiB | 4.915 | — |
| GLM-5.1 | 432.60 GiB | 4.929 | — |
| NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 | 334.55 GiB | 5.127 | — |
| ERNIE-Image-Turbo | 5.38 GiB | 5.755 | — |
●From the filewhat these mean