Model comparison

Qwen3-4B-Instruct vs gemma-4-E4B-it

At Q4_K_M, Qwen3-4B-Instruct is the smaller download — 2,497,278,784 bytes against 4,977,171,584.

From the file· summed bytes, KV per layer

Side by side

Qwen3-4B-Instructgemma-4-E4B-it
Parameters4.0B8.0B
Architectureqwen3gemma4
Layers42
Native context131,072
Mixture of expertsnono
Quantizations published836
Smallest quantization1.55 GiB3.30 GiB
Q4_K_M2.33 GiB4.64 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
ContextQwen3-4B-Instructgemma-4-E4B-itRatio
4,0960.12 GiB
8,1920.18 GiB
16,3840.29 GiB
32,7680.51 GiB
65,5360.94 GiB
131,0721.82 GiB