Quantization format
Q2_K_L
Q2_K_L has a nominal rate of — bits per weight. We have no measured files carrying this label yet.
From the file· 764 files measured
Nominal bpw
—
from the block layout
Measured average
3.379
764 files
Range
2.095–7.075
varies by architecture
File sizes
0.08 GiB+
up to 348.68 GiB
Real files
one per model, most downloaded first
| Model | Size● | Effective bpw● | vs nominal● |
|---|---|---|---|
| Qwen3.6-35B-A3B | 13.04 GiB | 3.116 | — |
| gemma-4-26B-A4B-it | 10.37 GiB | 3.355 | — |
| gemma-4-12B-it | 4.96 GiB | 3.560 | — |
| Qwen3.5-4B | 2.18 GiB | 4.027 | — |
| Hy3 | 99.81 GiB | 2.870 | — |
| Qwen3-Coder-30B-A3B-Instruct | 10.55 GiB | 2.969 | — |
| gemma-4-31B-it | 12.08 GiB | 3.319 | — |
| Qwen3-VL-30B-A3B-Instruct | 10.55 GiB | 2.918 | — |
| Llama-3.2-1B-Instruct | 0.54 GiB | 3.760 | — |
| gemma-4-E2B-it | 3.43 GiB | 5.756 | — |
| gpt-oss-20b | 10.95 GiB | 4.373 | — |
| Qwen3-8B | 3.19 GiB | 3.348 | — |
| Qwen3.5-0.8B | 0.49 GiB | 4.816 | — |
| Qwen3-4B | 1.55 GiB | 3.320 | — |
| Llama-3.1-8B-Instruct | 3.08 GiB | 3.290 | — |
| Ornith-1.0-35B | 12.21 GiB | 3.027 | — |
| Llama-3.2-3B-Instruct | 1.27 GiB | 3.396 | — |
| ThinkingCap-Qwen3.6-27B | 12.18 GiB | 3.825 | — |
| Qwythos-9B-v2 | 4.84 GiB | 4.305 | — |
| Qwen3.5-122B-A10B | 43.90 GiB | 3.015 | — |
| LFM2.5-1.2B-Instruct | 0.45 GiB | 3.304 | — |
| Qwen3-Coder-Next | 26.52 GiB | 2.859 | — |
| Qwen2.5-32B-Instruct | 12.18 GiB | 3.192 | — |
| gemma-3-1b-it | 0.64 GiB | 5.519 | — |
| Qwen3.5-35B-A3B | 13.04 GiB | 3.116 | — |
●From the filewhat these mean