Is deepseek-math-7b-instruct or Qwen3-8B smaller?

deepseek-math-7b-instruct is smaller at Q4_K_M: 3.93 GiB against 4.68 GiB.

Which uses less memory at long context?

Qwen3-8B — its KV cache at 32K context is 3.3x smaller, because of how its attention layers cache.

Model comparison

deepseek-math-7b-instruct vs Qwen3-8B

At Q4_K_M, deepseek-math-7b-instruct is the smaller download — 4,221,964,704 bytes against 5,027,783,488. At long context the gap widens: Qwen3-8B's KV cache at 32K is 3.3× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	deepseek-math-7b-instruct	Qwen3-8B
Parameters	6.9B	8.2B
Architecture	llama	qwen3
Layers	30	36
Native context	4,096	40,960
Mixture of experts	no	no
Quantizations published	38	51
Smallest quantization	2.53 GiB	2.12 GiB
Q4_K_M	3.93 GiB	4.68 GiB
Licence	other	apache-2.0

KV cache by context

the term that decides long-context viability

Context	deepseek-math-7b-instruct	Qwen3-8B	Ratio
4,096	1.88 GiB	0.56 GiB	3.33×
8,192	3.75 GiB	1.13 GiB	3.33×
16,384	7.50 GiB	2.25 GiB	3.33×
32,768	15.00 GiB	4.50 GiB	3.33×
65,536	30.00 GiB	9.00 GiB	3.33×
131,072	60.00 GiB	18.00 GiB	3.33×