Quantization format

I1_IQ3_M

I1_IQ3_M has a nominal rate of — bits per weight. We have no measured files carrying this label yet.

From the file· 1034 files measured

Nominal bpw

—

from the block layout

Measured average

3.796

1034 files

Range

2.094–8.066

varies by architecture

File sizes

0.09 GiB+

up to 160.39 GiB

Real files

one per model, most downloaded first

Model	Size●	Effective bpw●	vs nominal●
Qwopus3.6-35B-A3B-v1	14.38 GiB	3.436	—
Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTP	11.89 GiB	3.677	—
Qwen3-VL-8B-Instruct-abliterated-v1	3.63 GiB	3.556	—
Qwythos-9B-v2	4.21 GiB	3.748	—
gemma-4-26B-A4B-it-ultra-uncensored-heretic	11.54 GiB	3.842	—
Sugoi-14B-Ultra-HF	6.44 GiB	3.746	—
gemma-4-12b-heretic-abliterated	5.34 GiB	3.836	—
FastContext-1.0-4B-SFT	1.83 GiB	3.904	—
Qwen3.6-27B-uncensored-heretic-v2-Native-MTP-Preserved	11.89 GiB	3.734	—
gemma-4-31B-it-The-DECKARD-HERETIC-UNCENSORED-Thinking	13.43 GiB	3.690	—
Qwen3.6-35B-A3B-Claude-4.6-Opus-Reasoning-Distilled	14.38 GiB	3.436	—
Llama3.3-8B-Instruct-Thinking-Heretic-Uncensored-Claude-4.5-Opus-High-Reasoning	3.52 GiB	3.771	—
Huihui-ThinkingCap-Qwen3.6-27B-abliterated	11.89 GiB	3.734	—
Mistral-7B-Instruct-v0.2	3.06 GiB	3.629	—
Huihui-Qwen3-Coder-Next-abliterated	32.66 GiB	3.522	—
DeepSeek-R1-Distill-Llama-70B-abliterated	29.74 GiB	3.621	—
gemma-4-26B-A4B-it-heretic	11.54 GiB	3.842	—
gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-uncensored-heretic	5.34 GiB	3.836	—
Qwen3.5-122B-A10B-heretic-v2	50.09 GiB	3.524	—
Llama-3.3-70B-Instruct-abliterated	29.74 GiB	3.621	—
Qwable-9B-Claude-Fable-5	4.11 GiB	3.754	—
EVA-Qwen2.5-14B-v0.2	6.44 GiB	3.746	—
Prosopon-31B	14.00 GiB	3.679	—
Dirty-Muse-Writer-v01-Uncensored-Erotica-NSFW	4.19 GiB	3.891	—
Qwen3.6-34B-80L-Fable-5-Heretic	14.45 GiB	3.711	—

●From the filewhat these mean