Quantization publisher

anthracite-org

anthracite-org publishes 11 quantizations across 2 models in our index, averaging 5.478 effective bits per weight. Their files differ in size from other publishers' builds of the same nominal quantization on 8 of the pairs we can compare — the same label does not mean the same file.

From the file· summed file bytes
Repositories
2
Quantizations
11
Models covered
2
Avg effective bpw
5.478
across their files

Same model, same quant label, different bytes

largest disagreements first
ModelQuantanthracite-orgvsTheirsDifference
magnum-v2-12bIQ3_M5.33 GiBbartowski5.33 GiB-0.0%
magnum-v2-12bQ6_K9.37 GiBbartowski9.37 GiB-0.0%
magnum-v2-12bQ8_012.13 GiBbartowski12.13 GiB-0.0%
magnum-v2.5-12b-ktoQ5_K_M8.13 GiBbartowski8.13 GiB-0.0%
magnum-v2.5-12b-ktoIQ3_M5.33 GiBbartowski5.33 GiB-0.0%
magnum-v2.5-12b-ktoQ8_012.13 GiBbartowski12.13 GiB-0.0%
magnum-v2.5-12b-ktoIQ4_XS6.28 GiBbartowski6.28 GiB-0.0%
magnum-v2.5-12b-ktoQ4_K_M6.96 GiBbartowski6.96 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
magnum-v2.5-12b-kto65.18 GiB
magnum-v2-12b55.33 GiB