Quantization format

I1_IQ1_M

I1_IQ1_M has a nominal rate of — bits per weight. We have no measured files carrying this label yet.

From the file· 956 files measured

Nominal bpw

—

from the block layout

Measured average

2.209

956 files

Range

1.416–5.149

varies by architecture

File sizes

0.06 GiB+

up to 82.07 GiB

Real files

one per model, most downloaded first

Model	Size●	Effective bpw●	vs nominal●
Qwopus3.6-35B-A3B-v1	7.67 GiB	1.833	—
Qwen3-VL-8B-Instruct-abliterated-v1	2.10 GiB	2.059	—
gemma-4-26B-A4B-it-ultra-uncensored-heretic	8.07 GiB	2.687	—
Sugoi-14B-Ultra-HF	3.61 GiB	2.097	—
gemma-4-12b-heretic-abliterated	2.98 GiB	2.142	—
FastContext-1.0-4B-SFT	1.05 GiB	2.241	—
gemma-4-31B-it-The-DECKARD-HERETIC-UNCENSORED-Thinking	7.20 GiB	1.976	—
Qwen3.6-35B-A3B-Claude-4.6-Opus-Reasoning-Distilled	7.67 GiB	1.833	—
Llama3.3-8B-Instruct-Thinking-Heretic-Uncensored-Claude-4.5-Opus-High-Reasoning	2.01 GiB	2.154	—
Mistral-7B-Instruct-v0.2	1.63 GiB	1.938	—
Huihui-Qwen3-Coder-Next-abliterated	16.76 GiB	1.807	—
DeepSeek-R1-Distill-Llama-70B-abliterated	15.60 GiB	1.899	—
gemma-4-26B-A4B-it-heretic	8.07 GiB	2.687	—
gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-uncensored-heretic	2.98 GiB	2.142	—
Qwen3.5-122B-A10B-heretic-v2	25.70 GiB	1.808	—
Llama-3.3-70B-Instruct-abliterated	15.60 GiB	1.899	—
Qwable-9B-Claude-Fable-5	2.68 GiB	2.446	—
EVA-Qwen2.5-14B-v0.2	3.61 GiB	2.097	—
Prosopon-31B	7.63 GiB	2.004	—
Dirty-Muse-Writer-v01-Uncensored-Erotica-NSFW	2.37 GiB	2.204	—
Gemma-4-26B-A4B-StyleTune-V2	8.30 GiB	2.686	—
Qwen2.5-VL-7B-Instruct-abliterated	1.90 GiB	1.970	—
Qwen3.5-9B-Claude-4.6-Opus-Deckard-V4.2-Uncensored-Heretic-Thinking	2.68 GiB	2.446	—
Mistral-7B-Instruct-v0.1	1.63 GiB	1.938	—
Qwen3.5-9B-Claude-4.6-OS-Auto-Variable-HERETIC-UNCENSORED-THINKING	2.42 GiB	2.211	—

●From the filewhat these mean