community pool / model answer

gemma-3-text-4b-it-4bit

chat · mlx-community/gemma-3-text-4b-it-4bit

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

Reference rigQuantizationMedian resultBasisTTFT (ms)Peak VRAMMax contextEngines
M5 Max 128GB4-bit175.1 tok/sreported-3 GB2,048mlx
Pro 64GB4-bit105 tok/sreported1093 GB2,048mlx

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.