Quantization format

I1_Q2_K

I1_Q2_K has a nominal rate of — bits per weight. We have no measured files carrying this label yet.

From the file· 1037 files measured

Nominal bpw

—

from the block layout

Measured average

3.223

1037 files

Range

1.823–6.874

varies by architecture

File sizes

0.08 GiB+

up to 134.81 GiB

Real files

one per model, most downloaded first

Model	Size●	Effective bpw●	vs nominal●
Qwopus3.6-35B-A3B-v1	12.05 GiB	2.879	—
Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTP	10.12 GiB	3.129	—
Qwen3-VL-8B-Instruct-abliterated-v1	3.06 GiB	2.995	—
Qwythos-9B-v2	3.65 GiB	3.244	—
gemma-4-26B-A4B-it-ultra-uncensored-heretic	9.86 GiB	3.281	—
Sugoi-14B-Ultra-HF	5.37 GiB	3.126	—
gemma-4-12b-heretic-abliterated	4.50 GiB	3.231	—
FastContext-1.0-4B-SFT	1.55 GiB	3.320	—
Qwen3.6-27B-uncensored-heretic-v2-Native-MTP-Preserved	10.12 GiB	3.177	—
gemma-4-31B-it-The-DECKARD-HERETIC-UNCENSORED-Thinking	11.10 GiB	3.048	—
Qwen3.6-35B-A3B-Claude-4.6-Opus-Reasoning-Distilled	12.05 GiB	2.879	—
Llama3.3-8B-Instruct-Thinking-Heretic-Uncensored-Claude-4.5-Opus-High-Reasoning	2.96 GiB	3.167	—
Huihui-ThinkingCap-Qwen3.6-27B-abliterated	10.12 GiB	3.177	—
Mistral-7B-Instruct-v0.2	2.53 GiB	3.004	—
Huihui-Qwen3-Coder-Next-abliterated	27.13 GiB	2.925	—
DeepSeek-R1-Distill-Llama-70B-abliterated	24.56 GiB	2.991	—
gemma-4-26B-A4B-it-heretic	9.86 GiB	3.281	—
gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-uncensored-heretic	4.50 GiB	3.231	—
Qwen3.5-122B-A10B-heretic-v2	41.52 GiB	2.920	—
Llama-3.3-70B-Instruct-abliterated	24.56 GiB	2.991	—
Qwable-9B-Claude-Fable-5	3.56 GiB	3.254	—
EVA-Qwen2.5-14B-v0.2	5.37 GiB	3.126	—
Prosopon-31B	11.53 GiB	3.030	—
Dirty-Muse-Writer-v01-Uncensored-Erotica-NSFW	3.54 GiB	3.294	—
Qwen3.6-34B-80L-Fable-5-Heretic	12.27 GiB	3.151	—

●From the filewhat these mean