community pool / model answer

gemma-3-4b-it

chat · google/gemma-3-4b-it

112.1 tok/s · model median · 2 runs in the model record.

Reference rigQuantizationMedian resultBasisTTFT (ms)Peak VRAMMax contextEngines
M3 Ultra 512GB4-bit127.2 tok/sreported185- GB8,192llama.cpp
RTX 3060 12GB4-bit97 tok/sreported1,0634.9 GB4,608llama.cpp

Context tested

The table above is the tested context for each exact rig and quantization cell. Empty fields remain empty rather than being filled with an estimate.