Arao AI
Back to search results

Qwen2.5-Omni-3B · IQ4_XS

Start a discussion
AMD Radeon AI PRO R970032GBllama.cpp v0.41L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen2.5-Omni-3B (IQ4_XS) on AMD Radeon AI PRO R9700 with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.

183.61 tok/s

Decode

Decode speed

748.24 tok/s

Prefill

Prefill speed

1.06 s

TTFT

Time to first token

6.76 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

5 days ago

Last verified

Performance

  1. llama.cpp · IQ4_XS (current)Decode 183.61 · Prefill 748.24 ·
  2. llama.cpp · Q2_K Decode 223.75 · Prefill 955.8 ·
  3. Ollama · IQ2_XS Decode 228.57 · Prefill 1151.43 ·
  4. Ollama · IQ4_XS Decode 178.25 · Prefill 834.26 ·

Core figures

Decode (avg)
183.61 tok/s
Prefill (avg)
748.24 tok/s
TTFT (avg)
1.06 s
VRAM (avg)
6.76 GB
MTP acceptance rate
—
TTFB
1.47 GB
Power draw
266.9 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen2.5-Omni-3B
Quantization
IQ4_XS
Framework
llama.cpp
Version
v0.41
Context length
2048 tokens
Batch size
256
GPU layers
28
Flash Attention
On

Hardware

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

GPU
AMD Radeon AI PRO R9700
Nominal VRAM
32 GB
Measured VRAM (avg)
6.76 GB
OS
Ubuntu 22.04
Driver
amdgpu 6.10
CUDA
N/A
Power draw
266.9 W

Sources & evidence

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

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-13

    v0.41 · Ubuntu 22.04 · CUDA N/A · 2048 ctx

    183.61 tok/s

    Decode

    748.24 tok/s

    Prefill

    1.06 s

    TTFT

    6.76 GB

    VRAM

    —

    MTP

    266.9 W

    Power draw

    L2 batch seed 2026-10-13: llama.cpp on hw_id=79 running model_id=60 at IQ4_XS