Arao AI
Back to search results

Qwen3.5-0.8B · Q4_K_M

Start a discussion
AMD Ryzen AI Max+ 395128GBllama.cpp v0.39L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3.5-0.8B (Q4_K_M) on AMD Ryzen AI Max+ 395 with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.

68.16 tok/s

Decode

Decode speed

362.8 tok/s

Prefill

Prefill speed

0.56 s

TTFT

Time to first token

3.48 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 · IQ2_XS Decode 231.64 · Prefill 936.81 ·
  2. llama.cpp · Q4_K_M (current)Decode 68.16 · Prefill 362.8 ·
  3. vLLM · IQ4_XS Decode 85.62 · Prefill 384.79 ·
  4. llama.cpp · IQ4_XS Decode 273.83 · Prefill 1312.19 ·
  5. llama.cpp · Q5_K_M Decode 197.9 · Prefill 695.65 ·
  6. Ollama · Q4_K_M Decode 220.93 · Prefill 741.5 ·
  7. Ollama · Q8_0 Decode 163.22 · Prefill 720.03 ·
  8. llama.cpp · IQ1_S Decode 380.54 · Prefill 1677.44 ·

Core figures

Decode (avg)
68.16 tok/s
Prefill (avg)
362.8 tok/s
TTFT (avg)
0.56 s
VRAM (avg)
3.48 GB
MTP acceptance rate
—
TTFB
0.55 GB
Power draw
57 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3.5-0.8B
Quantization
Q4_K_M
Framework
llama.cpp
Version
v0.39
Context length
8192 tokens
Batch size
256
GPU layers
20
Flash Attention
Off

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)
3.48 GB
OS
Windows 11 23H2
Driver
amdgpu-pro 24.20
CUDA
12.6
Power draw
57 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

    v0.39 · Windows 11 23H2 · CUDA 12.6 · 8192 ctx

    68.16 tok/s

    Decode

    362.8 tok/s

    Prefill

    0.56 s

    TTFT

    3.48 GB

    VRAM

    —

    MTP

    57 W

    Power draw

    L2 batch seed 2026-10-19: llama.cpp on hw_id=24 running model_id=51 at Q4_K_M