Quantization format
UD_Q5_K_M
UD_Q5_K_M has a nominal rate of — bits per weight. We have no measured files carrying this label yet.
From the file· 31 files measured
Nominal bpw
—
from the block layout
Measured average
7.032
31 files
Range
5.157–15.896
varies by architecture
File sizes
5.92 GiB+
up to 705.94 GiB
Real files
one per model, most downloaded first
| Model | Size● | Effective bpw● | vs nominal● |
|---|---|---|---|
| Qwen3.6-35B-A3B | 24.64 GiB | 5.887 | — |
| gemma-4-26B-A4B-it | 19.70 GiB | 6.374 | — |
| Qwen-AgentWorld-35B-A3B | 24.64 GiB | 6.106 | — |
| LTX-2.3 | 16.79 GiB | 15.683 | — |
| GLM-5.2 | 522.31 GiB | 5.956 | — |
| Ornith-1.0-35B | 24.64 GiB | 6.106 | — |
| Qwen3.5-122B-A10B | 87.21 GiB | 5.989 | — |
| Qwen3-Coder-Next | 55.17 GiB | 5.948 | — |
| Laguna-S-2.1 | 81.83 GiB | 5.979 | — |
| Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16 | 27.00 GiB | 7.026 | — |
| Step-3.7-Flash | 136.43 GiB | 5.820 | — |
| LFM2.5-8B-A1B | 5.92 GiB | 6.010 | — |
| Ornith-1.0-397B | 275.01 GiB | 5.954 | — |
| Qwen3.5-397B-A17B | 273.50 GiB | 5.824 | — |
| North-Mini-Code-1.0 | 21.37 GiB | 6.022 | — |
| MiniMax-M2.7 | 157.23 GiB | 5.906 | — |
| MiniMax-M3 | 296.01 GiB | 5.954 | — |
| NVIDIA-Nemotron-3-Super-120B-A12B-BF16 | 99.96 GiB | 6.947 | — |
| Mistral-Small-4-119B-2603 | 83.04 GiB | 5.974 | — |
| Huihui-gemma-4-26B-A4B-it-abliterated | 19.70 GiB | 6.374 | — |
| MiMo-V2.5 | 214.37 GiB | 5.925 | — |
| GLM-5.1 | 520.11 GiB | 5.926 | — |
| NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 | 397.22 GiB | 6.087 | — |
| ERNIE-Image-Turbo | 6.27 GiB | 6.708 | — |
| MiMo-V2.5-Pro | 705.94 GiB | 5.926 | — |
●From the filewhat these mean