Is DeepSeek-V4-Flash or r1-1776 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 34.0x smaller, because of how its attention layers cache.

Model comparison

DeepSeek-V4-Flash vs r1-1776

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 34.0× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	DeepSeek-V4-Flash	r1-1776
Parameters	291B	671B
Architecture	deepseek4	deepseek2
Layers	43	61
Native context	1,048,576	163,840
Mixture of experts	yes, 256 experts	yes, 256 experts
Quantizations published	12	35
Smallest quantization	76.87 GiB	138.66 GiB
Q4_K_M	—	376.65 GiB
Licence	mit	mit

KV cache by context

the term that decides long-context viability

Context	DeepSeek-V4-Flash	r1-1776	Ratio
4,096	0.06 GiB	0.27 GiB	4.26×
8,192	0.06 GiB	0.54 GiB	8.51×
16,384	0.06 GiB	1.07 GiB	17.02×
32,768	0.06 GiB	2.14 GiB	34.05×
65,536	0.06 GiB	4.29 GiB	68.09×
131,072	0.06 GiB	8.58 GiB	136.19×