Quantization format

I1_IQ2_M

I1_IQ2_M has a nominal rate of — bits per weight. We have no measured files carrying this label yet.

From the file· 957 files measured

Nominal bpw

—

from the block layout

Measured average

2.950

957 files

Range

1.740–6.260

varies by architecture

File sizes

0.07 GiB+

up to 119.77 GiB

Real files

one per model, most downloaded first

Model	Size●	Effective bpw●	vs nominal●
Qwopus3.6-35B-A3B-v1	10.86 GiB	2.594	—
Qwen3-VL-8B-Instruct-abliterated-v1	2.84 GiB	2.785	—
gemma-4-26B-A4B-it-ultra-uncensored-heretic	9.66 GiB	3.217	—
Sugoi-14B-Ultra-HF	4.99 GiB	2.901	—
gemma-4-12b-heretic-abliterated	4.07 GiB	2.924	—
FastContext-1.0-4B-SFT	1.41 GiB	3.009	—
gemma-4-31B-it-The-DECKARD-HERETIC-UNCENSORED-Thinking	10.17 GiB	2.793	—
Qwen3.6-35B-A3B-Claude-4.6-Opus-Reasoning-Distilled	10.86 GiB	2.594	—
Llama3.3-8B-Instruct-Thinking-Heretic-Uncensored-Claude-4.5-Opus-High-Reasoning	2.75 GiB	2.937	—
Mistral-7B-Instruct-v0.2	2.33 GiB	2.763	—
Huihui-Qwen3-Coder-Next-abliterated	24.27 GiB	2.616	—
DeepSeek-R1-Distill-Llama-70B-abliterated	22.46 GiB	2.735	—
gemma-4-26B-A4B-it-heretic	9.67 GiB	3.217	—
gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-uncensored-heretic	4.07 GiB	2.924	—
Qwen3.5-122B-A10B-heretic-v2	37.20 GiB	2.616	—
Llama-3.3-70B-Instruct-abliterated	22.46 GiB	2.735	—
Qwable-9B-Claude-Fable-5	3.36 GiB	3.067	—
EVA-Qwen2.5-14B-v0.2	4.99 GiB	2.901	—
Prosopon-31B	10.73 GiB	2.821	—
Dirty-Muse-Writer-v01-Uncensored-Erotica-NSFW	3.20 GiB	2.973	—
Gemma-4-26B-A4B-StyleTune-V2	9.96 GiB	3.223	—
Qwen2.5-VL-7B-Instruct-abliterated	2.59 GiB	2.682	—
Qwen3.5-9B-Claude-4.6-Opus-Deckard-V4.2-Uncensored-Heretic-Thinking	3.36 GiB	3.067	—
MythoMax-L2-13b	4.41 GiB	2.909	—
Mistral-7B-Instruct-v0.1	2.33 GiB	2.763	—

●From the filewhat these mean