community pool / model answer

gemma-4-E4B-it-GGUF

chat · unsloth/gemma-4-E4B-it-GGUF

68.8 tok/s · model median · 35 runs in the model record.

Reference rig	Quantization	Median result	Basis	TTFT (ms)	Peak VRAM	Max context	Engines
Minisforum UM790 Pro 64GB	4-bit	21.7 tok/s	reported	-	- GB	2,048	llama.cpp
Radeon 890M 32GB	4-bit	18.7 tok/s	reported	1,350	- GB	2,048	llama.cpp
RTX 2080 Ti 11GB	4-bit	106 tok/s	reported	-	- GB	2,048	llama.cpp
RTX 2080 Ti 11GB	6-bit	89 tok/s	reported	-	- GB	2,048	llama.cpp
RTX 2080 Ti 11GB	8-bit	80 tok/s	reported	-	- GB	2,048	llama.cpp
RTX 3050 8GB	5-bit	46.4 tok/s	reported	169	- GB	128,000	llama.cpp
RTX 3060 12GB	4-bit	69.9 tok/s	measured	1,066	3.5 GB	131,072	llama.cpp
RTX 3060 12GB ×2	4-bit	69.3 tok/s	measured	5,185	5.3 GB	131,072	llama.cpp
Ryzen AI Max 385 24GB	4-bit	18.7 tok/s	reported	1,350	- GB	2,048	llama.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.