Is DeepSeek-V4-Flash or NVIDIA-Nemotron-3-Super-120B-A12B-BF16 smaller?

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

Which uses less memory at long context?

DeepSeek-V4-Flash — its KV cache at 32K context is 43.7x smaller, because of how its attention layers cache.

Model comparison

DeepSeek-V4-Flash vs NVIDIA-Nemotron-3-Super-120B-A12B-BF16

These two publish different quantization sets; the table below has the exact sizes. At long context the gap widens: DeepSeek-V4-Flash's KV cache at 32K is 43.7× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	DeepSeek-V4-Flash	NVIDIA-Nemotron-3-Super-120B-A12B-BF16
Parameters	291B	124B
Architecture	deepseek4	nemotron_h_moe
Layers	43	88
Native context	1,048,576	262,144
Mixture of experts	yes, 256 experts	yes, 512 experts
Quantizations published	12	45
Smallest quantization	76.87 GiB	43.19 GiB
Q4_K_M	—	80.14 GiB
Licence	mit	other

KV cache by context

the term that decides long-context viability

Context	DeepSeek-V4-Flash	NVIDIA-Nemotron-3-Super-120B-A12B-BF16	Ratio
4,096	0.06 GiB	0.34 GiB	5.46×
8,192	0.06 GiB	0.69 GiB	10.91×
16,384	0.06 GiB	1.38 GiB	21.83×
32,768	0.06 GiB	2.75 GiB	43.66×
65,536	0.06 GiB	5.50 GiB	87.32×
131,072	0.06 GiB	11.00 GiB	174.64×