Quantization format

I1_Q3_K_L

I1_Q3_K_L 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

4.321

1034 files

Range

1.151–9.199

varies by architecture

File sizes

0.10 GiB+

up to 148.65 GiB

Real files

one per model, most downloaded first

Model	Size●	Effective bpw●	vs nominal●
Qwopus3.6-35B-A3B-v1	16.87 GiB	4.031	—
Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTP	13.56 GiB	4.193	—
Qwen3-VL-8B-Instruct-abliterated-v1	4.13 GiB	4.044	—
Qwythos-9B-v2	4.70 GiB	4.184	—
gemma-4-26B-A4B-it-ultra-uncensored-heretic	12.88 GiB	4.286	—
Sugoi-14B-Ultra-HF	7.38 GiB	4.292	—
gemma-4-12b-heretic-abliterated	6.12 GiB	4.392	—
FastContext-1.0-4B-SFT	2.09 GiB	4.455	—
Qwen3.6-27B-uncensored-heretic-v2-Native-MTP-Preserved	13.56 GiB	4.258	—
gemma-4-31B-it-The-DECKARD-HERETIC-UNCENSORED-Thinking	15.49 GiB	4.254	—
Qwen3.6-35B-A3B-Claude-4.6-Opus-Reasoning-Distilled	16.87 GiB	4.031	—
Llama3.3-8B-Instruct-Thinking-Heretic-Uncensored-Claude-4.5-Opus-High-Reasoning	4.03 GiB	4.306	—
Huihui-ThinkingCap-Qwen3.6-27B-abliterated	13.56 GiB	4.258	—
Mistral-7B-Instruct-v0.2	3.56 GiB	4.222	—
Huihui-Qwen3-Coder-Next-abliterated	38.48 GiB	4.148	—
DeepSeek-R1-Distill-Llama-70B-abliterated	34.59 GiB	4.211	—
gemma-4-26B-A4B-it-heretic	12.88 GiB	4.286	—
gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-uncensored-heretic	6.12 GiB	4.392	—
Qwen3.5-122B-A10B-heretic-v2	58.85 GiB	4.140	—
Llama-3.3-70B-Instruct-abliterated	34.59 GiB	4.211	—
Qwable-9B-Claude-Fable-5	4.59 GiB	4.188	—
EVA-Qwen2.5-14B-v0.2	7.38 GiB	4.292	—
Prosopon-31B	16.05 GiB	4.218	—
Dirty-Muse-Writer-v01-Uncensored-Erotica-NSFW	4.78 GiB	4.443	—
Qwen3.6-34B-80L-Fable-5-Heretic	16.53 GiB	4.245	—

●From the filewhat these mean