Is HyperCLOVAX-SEED-Text-Instruct-0.5B or gemma-3-1b-it smaller?

HyperCLOVAX-SEED-Text-Instruct-0.5B is smaller at Q4_K_M: 0.40 GiB against 0.75 GiB.

Which uses less memory at long context?

gemma-3-1b-it — its KV cache at 32K context is 20.5x smaller, because of how its attention layers cache.

Model comparison

HyperCLOVAX-SEED-Text-Instruct-0.5B vs gemma-3-1b-it

At Q4_K_M, HyperCLOVAX-SEED-Text-Instruct-0.5B is the smaller download — 431,882,784 bytes against 806,058,240. At long context the gap widens: gemma-3-1b-it's KV cache at 32K is 20.5× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	HyperCLOVAX-SEED-Text-Instruct-0.5B	gemma-3-1b-it
Parameters	566M	1000M
Architecture	llama	gemma3
Layers	24	26
Native context	8,192	32,768
Mixture of experts	no	no
Quantizations published	4	28
Smallest quantization	0.39 GiB	0.52 GiB
Q4_K_M	0.40 GiB	0.75 GiB
Licence	other	gemma

KV cache by context

the term that decides long-context viability

Context	HyperCLOVAX-SEED-Text-Instruct-0.5B	gemma-3-1b-it	Ratio
4,096	0.38 GiB	0.04 GiB	10.11×
8,192	0.75 GiB	0.05 GiB	14.22×
16,384	1.50 GiB	0.08 GiB	17.86×
32,768	3.00 GiB	0.15 GiB	20.48×
65,536	6.00 GiB	0.27 GiB	22.10×
131,072	12.00 GiB	0.52 GiB	23.01×