Quantization format

I1_Q4_K_M

I1_Q4_K_M has a nominal rate of — bits per weight. We have no measured files carrying this label yet.

From the file· 1033 files measured

Nominal bpw

—

from the block layout

Measured average

4.925

1033 files

Range

1.238–10.460

varies by architecture

File sizes

0.11 GiB+

up to 173.97 GiB

Real files

one per model, most downloaded first

Model	Size●	Effective bpw●	vs nominal●
Qwopus3.6-35B-A3B-v1	19.71 GiB	4.710	—
Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTP	15.66 GiB	4.841	—
Qwen3-VL-8B-Instruct-abliterated-v1	4.68 GiB	4.588	—
Qwythos-9B-v2	5.38 GiB	4.790	—
gemma-4-26B-A4B-it-ultra-uncensored-heretic	15.64 GiB	5.207	—
Sugoi-14B-Ultra-HF	8.37 GiB	4.868	—
gemma-4-12b-heretic-abliterated	6.87 GiB	4.938	—
FastContext-1.0-4B-SFT	2.33 GiB	4.967	—
Qwen3.6-27B-uncensored-heretic-v2-Native-MTP-Preserved	15.66 GiB	4.916	—
gemma-4-31B-it-The-DECKARD-HERETIC-UNCENSORED-Thinking	17.40 GiB	4.780	—
Qwen3.6-35B-A3B-Claude-4.6-Opus-Reasoning-Distilled	19.71 GiB	4.710	—
Llama3.3-8B-Instruct-Thinking-Heretic-Uncensored-Claude-4.5-Opus-High-Reasoning	4.58 GiB	4.902	—
Huihui-ThinkingCap-Qwen3.6-27B-abliterated	15.66 GiB	4.916	—
Mistral-7B-Instruct-v0.2	4.07 GiB	4.826	—
Huihui-Qwen3-Coder-Next-abliterated	45.16 GiB	4.869	—
DeepSeek-R1-Distill-Llama-70B-abliterated	39.60 GiB	4.821	—
gemma-4-26B-A4B-it-heretic	15.64 GiB	5.207	—
gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-uncensored-heretic	6.87 GiB	4.938	—
Qwen3.5-122B-A10B-heretic-v2	69.11 GiB	4.862	—
Llama-3.3-70B-Instruct-abliterated	39.60 GiB	4.821	—
Qwable-9B-Claude-Fable-5	5.24 GiB	4.786	—
EVA-Qwen2.5-14B-v0.2	8.37 GiB	4.868	—
Prosopon-31B	18.14 GiB	4.768	—
Dirty-Muse-Writer-v01-Uncensored-Erotica-NSFW	5.37 GiB	4.987	—
Qwen3.6-34B-80L-Fable-5-Heretic	19.10 GiB	4.905	—

●From the filewhat these mean