Quantization format

TQ1_0

TQ1_0 has a nominal rate of bits per weight. We have no measured files carrying this label yet.

From the file· 14 files measured
Nominal bpw
from the block layout
Measured average
3.256
14 files
Range
1.811–10.411
varies by architecture
File sizes
0.02 GiB+
up to 92.59 GiB

Real files

one per model, most downloaded first
ModelSizeEffective bpwvs nominal
Qwen-Image-Edit-25096.50 GiB2.731
Qwen3-0.6B-Base0.29 GiB4.218
Llama-4-Maverick-17B-128E-Instruct92.59 GiB1.981
Wan_2.2_ComfyUI_Repackaged1.31 GiB
gemma-4-26B-A4B7.98 GiB2.583
Llama-3.2-3B0.86 GiB2.307
diffusiongemma-26B-A4B-it-HERETIC-Uncensored7.99 GiB2.657
Qwen3.5-9B-Base2.68 GiB2.388
bitnet-b1.58-2B-4T1.03 GiB10.411
granite-3.1-8b-instruct1.72 GiB1.811
Supra-Router-51M0.02 GiB3.884
Maestro1-9B2.17 GiB2.129
BugTraceAI-Apex-G4-26B-Master-f167.98 GiB2.717
Supra-50M-Base0.02 GiB3.884
granite-3.1-2b-instruct0.56 GiB1.889
Qwen3-Coder-Next-Opus-4.6-Reasoning-Distilled16.12 GiB
From the filewhat these mean

Other quantizations