community pool / model answer

gemma-3n-E4B-it

chat · google/gemma-3n-E4B-it

55 tok/s · model median · 4 runs in the model record.

Reference rigQuantizationMedian resultBasisTTFT (ms)Peak VRAMMax contextEngines
RTX 2080 8GB4-bit34 tok/sreported-7.5 GB2,048ollama
RTX 3080 12GB4-bit75 tok/sreported-8.2 GB2,048ollama
RTX3060+RTX2080-pooled-20GB 20GB4-bit74 tok/sreported-7.5 GB2,048ollama

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.