Quantization format
UD_IQ4_NL
UD_IQ4_NL has a nominal rate of — bits per weight. We have no measured files carrying this label yet.
From the file· 29 files measured
Nominal bpw
—
from the block layout
Measured average
4.012
29 files
Range
3.594–4.734
varies by architecture
File sizes
4.06 GiB+
up to 466.84 GiB
Real files
one per model, most downloaded first
| Model | Size● | Effective bpw● | vs nominal● |
|---|---|---|---|
| Qwen3.6-35B-A3B | 16.80 GiB | 4.014 | — |
| gemma-4-26B-A4B-it | 12.68 GiB | 4.103 | — |
| DeepSeek-V4-Flash | 128.43 GiB | 3.792 | — |
| Qwen-AgentWorld-35B-A3B | 16.87 GiB | 4.181 | — |
| GLM-5.2 | 347.07 GiB | 3.958 | — |
| Ornith-1.0-35B | 16.87 GiB | 4.181 | — |
| Qwen3.5-122B-A10B | 57.22 GiB | 3.929 | — |
| Qwen3-Coder-Next | 36.54 GiB | 3.939 | — |
| Qwen3.5-35B-A3B | 16.60 GiB | 3.966 | — |
| Laguna-S-2.1 | 54.71 GiB | 3.998 | — |
| Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16 | 18.19 GiB | 4.734 | — |
| Step-3.7-Flash | 90.63 GiB | 3.866 | — |
| LFM2.5-8B-A1B | 4.06 GiB | 4.118 | — |
| Ornith-1.0-397B | 182.60 GiB | 3.953 | — |
| Qwen3.5-397B-A17B | 180.45 GiB | 3.843 | — |
| North-Mini-Code-1.0 | 14.47 GiB | 4.076 | — |
| MiniMax-M2.7 | 103.15 GiB | 3.874 | — |
| MiniMax-M3 | 197.24 GiB | 3.967 | — |
| NVIDIA-Nemotron-3-Super-120B-A12B-BF16 | 60.06 GiB | 4.173 | — |
| Mistral-Small-4-119B-2603 | 55.13 GiB | 3.966 | — |
| Huihui-gemma-4-26B-A4B-it-abliterated | 12.50 GiB | 4.044 | — |
| MiMo-V2.5 | 142.12 GiB | 3.928 | — |
| GLM-5.1 | 343.54 GiB | 3.914 | — |
| NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 | 273.97 GiB | 4.199 | — |
| MiMo-V2.5-Pro | 466.84 GiB | 3.919 | — |
●From the filewhat these mean