Quantization format

I1_Q3_K_M

I1_Q3_K_M has a nominal rate of — bits per weight. We have no measured files carrying this label yet.

From the file· 1035 files measured

Nominal bpw

—

from the block layout

Measured average

4.029

1035 files

Range

1.061–8.593

varies by architecture

File sizes

0.09 GiB+

up to 176.30 GiB

Real files

one per model, most downloaded first

Model	Size●	Effective bpw●	vs nominal●
Qwopus3.6-35B-A3B-v1	15.61 GiB	3.731	—
Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTP	12.57 GiB	3.888	—
Qwen3-VL-8B-Instruct-abliterated-v1	3.84 GiB	3.763	—
Qwythos-9B-v2	4.41 GiB	3.926	—
gemma-4-26B-A4B-it-ultra-uncensored-heretic	12.37 GiB	4.119	—
Sugoi-14B-Ultra-HF	6.84 GiB	3.975	—
gemma-4-12b-heretic-abliterated	5.67 GiB	4.072	—
FastContext-1.0-4B-SFT	1.93 GiB	4.128	—
Qwen3.6-27B-uncensored-heretic-v2-Native-MTP-Preserved	12.57 GiB	3.948	—
gemma-4-31B-it-The-DECKARD-HERETIC-UNCENSORED-Thinking	14.24 GiB	3.911	—
Qwen3.6-35B-A3B-Claude-4.6-Opus-Reasoning-Distilled	15.61 GiB	3.731	—
Llama3.3-8B-Instruct-Thinking-Heretic-Uncensored-Claude-4.5-Opus-High-Reasoning	3.74 GiB	4.004	—
Huihui-ThinkingCap-Qwen3.6-27B-abliterated	12.57 GiB	3.948	—
Mistral-7B-Instruct-v0.2	3.28 GiB	3.888	—
Huihui-Qwen3-Coder-Next-abliterated	35.65 GiB	3.843	—
DeepSeek-R1-Distill-Llama-70B-abliterated	31.91 GiB	3.886	—
gemma-4-26B-A4B-it-heretic	12.37 GiB	4.119	—
gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-uncensored-heretic	5.67 GiB	4.072	—
Qwen3.5-122B-A10B-heretic-v2	54.58 GiB	3.839	—
Llama-3.3-70B-Instruct-abliterated	31.91 GiB	3.886	—
Qwable-9B-Claude-Fable-5	4.31 GiB	3.931	—
EVA-Qwen2.5-14B-v0.2	6.84 GiB	3.975	—
Prosopon-31B	14.80 GiB	3.890	—
Dirty-Muse-Writer-v01-Uncensored-Erotica-NSFW	4.43 GiB	4.122	—
Qwen3.6-34B-80L-Fable-5-Heretic	15.29 GiB	3.928	—

●From the filewhat these mean