Quantization format
Q5_0
Q5_0 has a nominal rate of — bits per weight. We have no measured files carrying this label yet.
From the file· 506 files measured
Nominal bpw
—
from the block layout
Measured average
6.111
506 files
Range
1.558–31.893
varies by architecture
File sizes
0.02 GiB+
up to 151.16 GiB
Real files
one per model, most downloaded first
| Model | Size● | Effective bpw● | vs nominal● |
|---|---|---|---|
| FLUX.2-klein-9B | 6.25 GiB | 5.910 | — |
| cohere-transcribe-03-2026 | 1.62 GiB | 6.733 | — |
| LTX-2.3 | 14.24 GiB | 13.297 | — |
| Qwen3-8B | 5.33 GiB | 5.588 | — |
| Qwen3-4B | 2.63 GiB | 5.616 | — |
| parakeet-tdt-0.6b-v3 | 0.44 GiB | 6.000 | — |
| whisper-medium | 0.50 GiB | 5.647 | — |
| Wan2.2-I2V-A14B | 9.61 GiB | 5.775 | — |
| Wan2.1-T2V-1.3B | 0.97 GiB | 5.861 | — |
| Qwen2.5-32B-Instruct | 21.08 GiB | 5.528 | — |
| Qwen2.5-Coder-7B-Instruct | 4.95 GiB | 5.583 | — |
| Qwen3-14B | 9.56 GiB | 5.560 | — |
| Qwen2.5-7B-Instruct | 4.95 GiB | 5.583 | — |
| Jan-v3-4B-base-instruct | 2.88 GiB | 5.606 | — |
| Qwen2.5-1.5B-Instruct | 1.17 GiB | 6.525 | — |
| whisper-large-v3 | 1.01 GiB | 5.604 | — |
| Qwen3-30B-A3B | 19.63 GiB | 5.524 | — |
| LTX-2 | 12.72 GiB | 5.791 | — |
| Qwen2.5-Coder-32B-Instruct | 42.17 GiB | 11.055 | — |
| whisper-large-v3-turbo | 0.53 GiB | 5.677 | — |
| Qwen2.5-Coder-14B-Instruct | 19.12 GiB | 11.121 | — |
| granite-4.1-3b | 2.21 GiB | 5.590 | — |
| Qwen2.5-3B-Instruct | 2.22 GiB | 6.179 | — |
| Qwen2.5-0.5B-Instruct | 0.46 GiB | 7.942 | — |
| Qwen3-32B | 21.08 GiB | 5.527 | — |
●From the filewhat these mean