Quantization publisher

TrevorJS

TrevorJS publishes 13 quantizations across 6 models in our index, averaging 7.157 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
6
Quantizations
13
Models covered
6
Avg effective bpw
7.157
across their files

Same model, same quant label, different bytes

largest disagreements first
ModelQuantTrevorJSvsTheirsDifference
gemma-4-E2B-it-uncensoredQ8_04.63 GiBHandyfff3.99 GiB+15.8%
gemma-4-E2B-it-uncensoredQ4_K_M3.19 GiBHandyfff2.70 GiB+18.1%
gemma-4-E4B-it-uncensoredQ4_K_M4.97 GiBlowlevelgoblin4.97 GiB0.0%
gemma-4-E2B-it-uncensoredQ4_K_M3.19 GiBmradermacher3.19 GiB-0.0%
gemma-4-E2B-it-uncensoredQ8_04.63 GiBmradermacher4.63 GiB-0.0%
Qwen3-Omni-30B-A3B-InstructQ4_K_M17.28 GiBggml-org17.28 GiB0.0%
Qwen3-Omni-30B-A3B-InstructQ8_030.25 GiBggml-org30.25 GiB0.0%

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
Qwen3-Omni-30B-A3B-Instruct317.28 GiB
gemma-4-26B-A4B-it-uncensored215.64 GiB
gemma-4-31B-it-uncensored217.40 GiB
gemma-4-E4B-it-uncensored24.97 GiB
gemma-4-12B-it-uncensored26.87 GiB
gemma-4-E2B-it-uncensored23.19 GiB