Is DeepSeek-V4-Flash or dots.llm1.inst 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 492.2x smaller, because of how its attention layers cache.

Model comparison

DeepSeek-V4-Flash vs dots.llm1.inst

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

From the file· summed bytes, KV per layer

Side by side

	DeepSeek-V4-Flash	dots.llm1.inst
Parameters	291B	143B
Architecture	deepseek4	dots1
Layers	43	62
Native context	1,048,576	32,768
Mixture of experts	yes, 256 experts	yes, 128 experts
Quantizations published	12	46
Smallest quantization	76.87 GiB	44.04 GiB
Q4_K_M	—	88.00 GiB
Licence	mit	mit

KV cache by context

the term that decides long-context viability

Context	DeepSeek-V4-Flash	dots.llm1.inst	Ratio
4,096	0.06 GiB	3.88 GiB	61.52×
8,192	0.06 GiB	7.75 GiB	123.04×
16,384	0.06 GiB	15.50 GiB	246.08×
32,768	0.06 GiB	31.00 GiB	492.16×
65,536	0.06 GiB	62.00 GiB	984.31×
131,072	0.06 GiB	124.00 GiB	1968.62×