Arao AI
Back to search results

DeepSeek-R1-Distill-Qwen-14B · Q4_K_S

Start a discussion
AMD Ryzen AI Max+ 395128GBvLLM v0.26L2 Cross-framework verified

This page aggregates 1 real-world runs of DeepSeek-R1-Distill-Qwen-14B (Q4_K_S) on AMD Ryzen AI Max+ 395 with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.

29.55 tok/s

Decode

Decode speed

96.92 tok/s

Prefill

Prefill speed

2.13 s

TTFT

Time to first token

25.03 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

17 days ago

Last verified

Performance

  1. vLLM · Q3_K_S Decode 34.77 · Prefill 161.9 ·
  2. vLLM · Q4_K_S (current)Decode 29.55 · Prefill 96.92 ·
  3. Ollama · Q4_K_S Decode 28.23 · Prefill 99.22 ·
  4. Ollama · Q3_K_M Decode 77.2 · Prefill 347.81 ·

Core figures

Decode (avg)
29.55 tok/s
Prefill (avg)
96.92 tok/s
TTFT (avg)
2.13 s
VRAM (avg)
25.03 GB
MTP acceptance rate
—
TTFB
0.24 GB
Power draw
49.7 W

Configuration

Member-level fields are taken from the most recent run

Model
DeepSeek-R1-Distill-Qwen-14B
Quantization
Q4_K_S
Framework
vLLM
Version
v0.26
Context length
16384 tokens
Batch size
8
GPU layers
999
Flash Attention
On

Hardware

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

GPU
AMD Ryzen AI Max+ 395
Nominal VRAM
128 GB
Measured VRAM (avg)
25.03 GB
OS
Windows 11 23H2
Driver
amdgpu 6.12
CUDA
12.6
Power draw
49.7 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

    v0.26 · Windows 11 23H2 · CUDA 12.6 · 16384 ctx

    29.55 tok/s

    Decode

    96.92 tok/s

    Prefill

    2.13 s

    TTFT

    25.03 GB

    VRAM

    —

    MTP

    49.7 W

    Power draw

    L2 batch seed 2026-10-19: vLLM on hw_id=24 running model_id=71 at Q4_K_S