Rent a GPU

Can the Apple M3 Ultra (256GB) run DeepSeek V4 Flash 284B-A13B?

Apple M3 Ultra (256GB) × DeepSeek V4 Flash 284B-A13B8K context
YesRuns fully on the GPU
Memory at Q4_K_M173 GB of 192 GB
Weights 172.2 GBKV cache 0.2 GBRuntime 0.6 GB
QuantQ4_K_MGood, most popular
Speed~62tokens per second
Max context740Ktokens that fit at Q4_K_M
  1. FP16
  2. Q8_0
  3. Q6_K
  4. Q5_K_M
  5. Q4_K_M
  6. Q3_K_M
  7. Q2_K

Yes. The Apple M3 Ultra (256GB) can run DeepSeek V4 Flash 284B-A13B fully in unified memory at Q4_K_M (good, most popular), using about 173 GB of the 192 GB available, at roughly 62 tokens/second.

Try other context lengths

Run it with Ollama

With Ollama installed, one command downloads and starts DeepSeek V4 Flash 284B-A13B:

ollama run deepseek-v4-flash

Ollama's default download is Q4_K_M (about 173 GB here). That is also the best quant that fits.

New to this? Step-by-step setup for the Apple M3 Ultra (256GB)

Run it in the cloud

No Apple M3 Ultra (256GB)? Run DeepSeek V4 Flash 284B-A13B in the cloud.

Rent a cloud GPU by the hour. No upfront cost, stop any time.

Rent a GPU on RunPodRent a GPU on Vast.ai

Memory needed by quantization

At 8,192 tokens of context. The black line marks the 192 GB the Apple M3 Ultra (256GB) can use. macOS lets the GPU use about 192 GB of the 256 GB unified memory by default.

QuantQualityMemoryFits?Speed
FP16Full quality568.8 GBToo big—
Q8_0Near-lossless302.6 GBToo big—
Q6_KExcellent233.7 GBToo big—
Q5_K_MVery good202.8 GBToo big—
Q4_K_MGood, most popular173 GBFits~62 tok/s
Q3_K_MNoticeably degraded139.6 GBFits~77 tok/s
Q2_KHeavily degraded107.3 GBFits~101 tok/s

Measured speed from real Apple M3 Ultra (256GB) owners

Tokens per second that people measured on their own Apple M3 Ultra (256GB), next to our estimate.

No one has measured DeepSeek V4 Flash 284B-A13B on the Apple M3 Ultra (256GB) yet. The speeds above are estimates. Run the benchmark and yours becomes the first real number on this page.

Have an Apple M3 Ultra (256GB)? Add your real speed

One command times DeepSeek V4 Flash 284B-A13B in your own Ollama, shows you exactly what it would send, and asks before sending anything. No sign-up, no personal data.

macOS and Linux

curl -fsSL https://gpufit.dev/bench.sh | sh -s -- deepseek-v4-flash

Windows (PowerShell)

$env:GPUFIT_MODEL="deepseek-v4-flash"; irm https://gpufit.dev/bench.ps1 | iex

How the benchmark works

Models that run well on the Apple M3 Ultra (256GB)

Frequently asked questions

Can Apple M3 Ultra (256GB) run DeepSeek V4 Flash 284B-A13B?

Yes. The Apple M3 Ultra (256GB) can run DeepSeek V4 Flash 284B-A13B fully in unified memory at Q4_K_M (good, most popular), using about 173 GB of the 192 GB available, at roughly 62 tokens/second.

How much VRAM does DeepSeek V4 Flash 284B-A13B need?

About 173 GB at Q4_K_M with 8,192 tokens of context, and about 568.8 GB at full FP16 precision.

How fast is DeepSeek V4 Flash 284B-A13B on the Apple M3 Ultra (256GB)?

Roughly 62 tokens per second at Q4_K_M, estimated from the Apple M3 Ultra (256GB)'s 819 GB/s memory bandwidth.