Giant tier

DeepSeek V4 Flash

DeepSeek · DeepSeek

The faster, more practical sibling of the 1.6-trillion-parameter DeepSeek V4 Pro. At 284B total it activates only ~13B parameters per token, so once loaded it generates far faster than its size suggests — but even at 4-bit it needs roughly 180 GB of memory, putting it in 256 GB+ unified-memory Mac or multi-GPU server territory. MIT licensed, with a 1M-token context window.

Params

284B (MoE, ~13B active)

License

MIT

Context

1M

Min VRAM

~180 GB (Q4)

Min RAM

256 GB

Tier

Giant

Run locally with Ollama

terminal
ollama run deepseek-v4-flash

Once running, Ollama exposes an OpenAI-compatible endpoint at localhost:11434/v1.

Will it run on my hardware?

DeepSeek V4 Flash needs about ~180 GB (Q4) of fast memory at Q4. Here is what in our hardware database clears that bar.

None of our tracked consumer hardware has enough memory for this model — it is server / datacenter territory.

15 other tracked configurations do not have enough memory.

Hardware links are affiliate links. We earn a small commission if you buy through them — at no extra cost to you. Disclosure

Specifications

Parameters284B (MoE, ~13B active)
Active params (MoE)~13B active per token
Context window1M
LicenseMIT
Min VRAM (Q4)~180 GB (Q4)
Min RAM256 GB
Recommended tierGiant
Best forcodinganalysisreasoningchat

Variants on Ollama

Ollama serves the Q4_K_M quantization by default. Higher-precision tags exist for users with more memory.

DeepSeek V4 Flash

default · Q4_K_M
ollama run deepseek-v4-flash
higher precision · q8_0 (generic tag)
ollama pull deepseek-v4-flash:q8_0
full precision · fp16 (generic tag)
ollama pull deepseek-v4-flash:fp16

Exact quantization tags vary per model — run ollama show deepseek-v4-flash or check the tags page for the full list.

Popularity trend (illustrative)

Illustrative adoption trend, not actual download counts. We do not publish a download number for any model because we have no honest source for it.

Similar models