Is Qwen2-1.5B-Instruct 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 6.1x smaller, because of how its attention layers cache.

Model comparison

Qwen2-1.5B-Instruct 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 6.1× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	Qwen2-1.5B-Instruct	embeddinggemma-300m
Parameters	1.5B	303M
Architecture	qwen2	gemma-embedding
Layers	28	24
Native context	32,768	2,048
Mixture of experts	no	no
Quantizations published	37	10
Smallest quantization	0.41 GiB	0.26 GiB
Q4_K_M	0.92 GiB	—
Licence	apache-2.0	—

KV cache by context

the term that decides long-context viability

Context	Qwen2-1.5B-Instruct	embeddinggemma-300m	Ratio
4,096	0.11 GiB	0.04 GiB	3.11×
8,192	0.22 GiB	0.05 GiB	4.31×
16,384	0.44 GiB	0.08 GiB	5.33×
32,768	0.88 GiB	0.14 GiB	6.05×
65,536	1.75 GiB	0.27 GiB	6.49×
131,072	3.50 GiB	0.52 GiB	6.74×