Is Qwen3.8-27B or Llama-3_3-Nemotron-Super-49B-v1_5 smaller?

Qwen3.8-27B is smaller at Q4_K_M: 15.66 GiB against 28.14 GiB.

Which uses less memory at long context?

Qwen3.8-27B — its KV cache at 32K context is 40.0x smaller, because of how its attention layers cache.

Model comparison

Qwen3.8-27B vs Llama-3_3-Nemotron-Super-49B-v1_5

At Q4_K_M, Qwen3.8-27B is the smaller download — 16,810,714,336 bytes against 30,215,578,432. At long context the gap widens: Qwen3.8-27B's KV cache at 32K is 40.0× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	Qwen3.8-27B	Llama-3_3-Nemotron-Super-49B-v1_5
Parameters	27.8B	49.9B
Architecture	qwen35	deci
Layers	64	80
Native context	262,144	131,072
Mixture of experts	no	no
Quantizations published	22	53
Smallest quantization	8.39 GiB	10.66 GiB
Q4_K_M	15.66 GiB	28.14 GiB
Licence	apache-2.0	—

KV cache by context

the term that decides long-context viability

Context	Qwen3.8-27B	Llama-3_3-Nemotron-Super-49B-v1_5	Ratio
4,096	0.25 GiB	10.00 GiB	39.97×
8,192	0.50 GiB	20.00 GiB	39.99×
16,384	1.00 GiB	40.00 GiB	39.99×
32,768	2.00 GiB	80.00 GiB	40.00×
65,536	4.00 GiB	160.00 GiB	40.00×
131,072	8.00 GiB	320.00 GiB	40.00×