Model comparison

medgemma-4b-it vs Qwen3-VL-2B-Instruct

At Q4_K_M, Qwen3-VL-2B-Instruct is the smaller download — 1,107,409,952 bytes against 2,489,893,664. At long context the gap widens: medgemma-4b-it's KV cache at 32K is 4.4× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

medgemma-4b-itQwen3-VL-2B-Instruct
Parameters4.3B2.1B
Architecturegemma3qwen3vl
Layers3428
Native context131,072262,144
Mixture of expertsnono
Quantizations published4646
Smallest quantization1.10 GiB0.50 GiB
Q4_K_M2.32 GiB1.03 GiB
Licenceotherapache-2.0

KV cache by context

the term that decides long-context viability
Contextmedgemma-4b-itQwen3-VL-2B-InstructRatio
4,0960.25 GiB0.44 GiB1.76×
8,1920.33 GiB0.88 GiB2.68×
16,3840.48 GiB1.75 GiB3.63×
32,7680.79 GiB3.50 GiB4.40×
65,5361.42 GiB7.00 GiB4.93×
131,0722.67 GiB14.00 GiB5.24×