Is granite-4.0-1b or embeddinggemma-300m smaller?

Both publish quantizations; compare the table for the exact sizes.

Which uses less memory at long context?

embeddinggemma-300m — its KV cache at 32K context is 17.3x smaller, because of how its attention layers cache.

Model comparison

granite-4.0-1b vs embeddinggemma-300m

These two publish different quantization sets; the table below has the exact sizes. At long context the gap widens: embeddinggemma-300m's KV cache at 32K is 17.3× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	granite-4.0-1b	embeddinggemma-300m
Parameters	1.6B	303M
Architecture	granite	gemma-embedding
Layers	40	24
Native context	131,072	2,048
Mixture of experts	no	no
Quantizations published	35	10
Smallest quantization	0.44 GiB	0.26 GiB
Q4_K_M	0.95 GiB	—
Licence	apache-2.0	—

KV cache by context

the term that decides long-context viability

Context	granite-4.0-1b	embeddinggemma-300m	Ratio
4,096	0.31 GiB	0.04 GiB	8.89×
8,192	0.63 GiB	0.05 GiB	12.31×
16,384	1.25 GiB	0.08 GiB	15.24×
32,768	2.50 GiB	0.14 GiB	17.30×
65,536	5.00 GiB	0.27 GiB	18.55×
131,072	10.00 GiB	0.52 GiB	19.25×