Is LCO-Embedding-Omni-3B-2605 or Qwen3-Embedding-8B smaller?

LCO-Embedding-Omni-3B-2605 is smaller at Q4_K_M: 1.96 GiB against 4.36 GiB.

Which uses less memory at long context?

LCO-Embedding-Omni-3B-2605 — its KV cache at 32K context is 4.0x smaller, because of how its attention layers cache.

Model comparison

LCO-Embedding-Omni-3B-2605 vs Qwen3-Embedding-8B

At Q4_K_M, LCO-Embedding-Omni-3B-2605 is the smaller download — 2,104,931,744 bytes against 4,676,804,928. At long context the gap widens: LCO-Embedding-Omni-3B-2605's KV cache at 32K is 4.0× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	LCO-Embedding-Omni-3B-2605	Qwen3-Embedding-8B
Parameters	4.7B	7.6B
Architecture	qwen2vl	qwen3
Layers	36	36
Native context	32,768	40,960
Mixture of experts	no	no
Quantizations published	4	17
Smallest quantization	1.96 GiB	2.87 GiB
Q4_K_M	1.96 GiB	4.36 GiB
Licence	apache-2.0	apache-2.0

KV cache by context

the term that decides long-context viability

Context	LCO-Embedding-Omni-3B-2605	Qwen3-Embedding-8B	Ratio
4,096	0.14 GiB	0.56 GiB	4.00×
8,192	0.28 GiB	1.13 GiB	4.00×
16,384	0.56 GiB	2.25 GiB	4.00×
32,768	1.13 GiB	4.50 GiB	4.00×
65,536	2.25 GiB	9.00 GiB	4.00×
131,072	4.50 GiB	18.00 GiB	4.00×