Arao AI
Back to search results

Qwen3.5-4B · FP16

Start a discussion
AMD Ryzen AI Max+ 395128GBOllama v0.22L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3.5-4B (FP16) on AMD Ryzen AI Max+ 395 with Ollama, contributed by 1 independent source platforms; metrics are averages of published measurements.

19.79 tok/s

Decode

Decode speed

92.05 tok/s

Prefill

Prefill speed

1.96 s

TTFT

Time to first token

17.64 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 · IQ2_M Decode 61.41 · Prefill 198.56 ·
  2. Ollama · FP16 (current)Decode 19.79 · Prefill 92.05 ·
  3. llama.cpp · IQ3_XS Decode 156.15 · Prefill 562.48 ·
  4. vLLM · IQ2_XS Decode 205.56 · Prefill 924.99 ·
  5. vLLM · Q6_K Decode 110.75 · Prefill 500.4 ·

Core figures

Decode (avg)
19.79 tok/s
Prefill (avg)
92.05 tok/s
TTFT (avg)
1.96 s
VRAM (avg)
17.64 GB
MTP acceptance rate
—
TTFB
0.16 GB
Power draw
50 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3.5-4B
Quantization
FP16
Framework
Ollama
Version
v0.22
Context length
8192 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
AMD Ryzen AI Max+ 395
Nominal VRAM
128 GB
Measured VRAM (avg)
17.64 GB
OS
Ubuntu 22.04
Driver
amdgpu-pro 24.20
CUDA
12.6
Power draw
50 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

    v0.22 · Ubuntu 22.04 · CUDA 12.6 · 8192 ctx

    19.79 tok/s

    Decode

    92.05 tok/s

    Prefill

    1.96 s

    TTFT

    17.64 GB

    VRAM

    —

    MTP

    50 W

    Power draw

    L2 batch seed 2026-10-19: Ollama on hw_id=24 running model_id=48 at FP16