Arao AI
Back to search results

Qwen2.5-0.5B-Instruct · IQ3_XS

Start a discussion
AMD Radeon AI PRO R970032GBvLLM v0.32L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen2.5-0.5B-Instruct (IQ3_XS) on AMD Radeon AI PRO R9700 with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.

346.96 tok/s

Decode

Decode speed

1,652.99 tok/s

Prefill

Prefill speed

1 s

TTFT

Time to first token

2.75 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

5 days ago

Last verified

Performance

  1. vLLM · IQ3_XS (current)Decode 346.96 · Prefill 1652.99 ·
  2. vLLM · Q5_K_S Decode 310.96 · Prefill 1151.48 ·
  3. Strata · Q3_K_M Decode 294.81 · Prefill 1272.19 ·
  4. Strata · Q5_K_M Decode 384.75 · Prefill 2096.67 ·

Core figures

Decode (avg)
346.96 tok/s
Prefill (avg)
1,652.99 tok/s
TTFT (avg)
1 s
VRAM (avg)
2.75 GB
MTP acceptance rate
—
TTFB
2.78 GB
Power draw
315.9 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen2.5-0.5B-Instruct
Quantization
IQ3_XS
Framework
vLLM
Version
v0.32
Context length
32768 tokens
Batch size
32
GPU layers
999
Flash Attention
Off

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)
2.75 GB
OS
Ubuntu 24.04
Driver
amdgpu-pro 24.20
CUDA
N/A
Power draw
315.9 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

    v0.32 · Ubuntu 24.04 · CUDA N/A · 32768 ctx

    346.96 tok/s

    Decode

    1,652.99 tok/s

    Prefill

    1 s

    TTFT

    2.75 GB

    VRAM

    —

    MTP

    315.9 W

    Power draw

    L2 batch seed 2026-10-13: vLLM on hw_id=79 running model_id=59 at IQ3_XS