Arao AI
Back to search results

DeepSeek-R1-Distill-Qwen-14B · Q5_K_M

Start a discussion
Apple M5 Ultra512GBllama.cpp v0.30L2 Cross-framework verified

This page aggregates 1 real-world runs of DeepSeek-R1-Distill-Qwen-14B (Q5_K_M) on Apple M5 Ultra with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.

77.61 tok/s

Decode

Decode speed

409.9 tok/s

Prefill

Prefill speed

2 s

TTFT

Time to first token

29.44 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

17 days ago

Last verified

Performance

  1. llama.cpp · Q5_K_M (current)Decode 77.61 · Prefill 409.9 ·
  2. Ollama · Q2_K Decode 118.32 · Prefill 453.53 ·
  3. Ollama · Q5_K_M Decode 65.66 · Prefill 313.69 ·

Core figures

Decode (avg)
77.61 tok/s
Prefill (avg)
409.9 tok/s
TTFT (avg)
2 s
VRAM (avg)
29.44 GB
MTP acceptance rate
—
TTFB
0.62 GB
Power draw
121.7 W

Configuration

Member-level fields are taken from the most recent run

Model
DeepSeek-R1-Distill-Qwen-14B
Quantization
Q5_K_M
Framework
llama.cpp
Version
v0.30
Context length
32768 tokens
Batch size
256
GPU layers
33
Flash Attention
On

Hardware

Nominal and measured figures are shown side by side; whether it runs is the reader's call

GPU
Apple M5 Ultra
Nominal VRAM
512 GB
Measured VRAM (avg)
29.44 GB
OS
macOS 26.1
Power draw
121.7 W

Sources & evidence

1 measured records in total, each traceable to its original source

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-19

    v0.30 · macOS 26.1 · 32768 ctx

    77.61 tok/s

    Decode

    409.9 tok/s

    Prefill

    2 s

    TTFT

    29.44 GB

    VRAM

    —

    MTP

    121.7 W

    Power draw

    L2 batch seed 2026-10-19: llama.cpp on hw_id=29 running model_id=71 at Q5_K_M