Quantization publisher

MRockatansky

MRockatansky publishes 15 quantizations across 5 models in our index, averaging 8.552 effective bits per weight. Their files differ in size from other publishers' builds of the same nominal quantization on 7 of the pairs we can compare — the same label does not mean the same file.

From the file· summed file bytes
Repositories
5
Quantizations
15
Models covered
5
Avg effective bpw
8.552
across their files

Same model, same quant label, different bytes

largest disagreements first
ModelQuantMRockatanskyvsTheirsDifference
gemma-4-31B-MergemaxxedIQ4_XS15.59 GiBmradermacher15.70 GiB-0.7%
gemma-4-31B-anthologyIQ4_XS15.59 GiBmradermacher15.70 GiB-0.7%
gemma-4-31B-anthologyQ4_K_M17.40 GiBmradermacher17.40 GiB-0.0%
gemma-4-31B-anthologyQ5_K_S19.85 GiBmradermacher19.85 GiB-0.0%
Cogidonia-v2-24BIQ4_XS12.00 GiBmradermacher12.00 GiB-0.0%
Cogidonia-v2-24BQ4_K_M13.35 GiBmradermacher13.35 GiB-0.0%
Cogidonia-v2-24BQ6_K18.02 GiBmradermacher18.02 GiB-0.0%

A quantization label describes a target, not a recipe. Publishers make different choices about which tensors to keep at higher precision, and some apply an importance matrix while others don't — so two files both honestly labelled the same thing can differ measurably in size and in quality.

Models they publish

ModelQuantizationsSmallest
gemma-4-31B-anthology515.59 GiB
gemma-4-31B-Mergemaxxed315.59 GiB
Cogidonia-v2-24B412.00 GiB
Gemma-4-31B-Storymaxxed3217.72 GiB
Gemma-4-31B-storymaxxed2157.20 GiB