Is grok-2 or DeepSeek-V4-Flash 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 127.0x smaller, because of how its attention layers cache.

Model comparison

grok-2 vs DeepSeek-V4-Flash

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

From the file· summed bytes, KV per layer

Side by side

	grok-2	DeepSeek-V4-Flash
Parameters	270B	291B
Architecture	grok	deepseek4
Layers	64	43
Native context	131,072	1,048,576
Mixture of experts	yes, 8 experts	yes, 256 experts
Quantizations published	66	12
Smallest quantization	57.16 GiB	76.87 GiB
Q4_K_M	152.79 GiB	—
Licence	other	mit

KV cache by context

the term that decides long-context viability

Context	grok-2	DeepSeek-V4-Flash	Ratio
4,096	1.00 GiB	0.06 GiB	15.88×
8,192	2.00 GiB	0.06 GiB	31.75×
16,384	4.00 GiB	0.06 GiB	63.50×
32,768	8.00 GiB	0.06 GiB	127.01×
65,536	16.00 GiB	0.06 GiB	254.02×
131,072	32.00 GiB	0.06 GiB	508.03×