Arao AI
Back to search results

Qwen3-Coder-Next · Q8_0

Start a discussion
2x AMD Strix HalovLLM v0.40L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3-Coder-Next (Q8_0) on 2x AMD Strix Halo with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.

87.69 tok/s

Decode

Decode speed

408.22 tok/s

Prefill

Prefill speed

0.88 s

TTFT

Time to first token

1.05 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

2 days ago

Last verified

Performance

  1. vLLM · Q3_K_S Decode 149.19 · Prefill 617.1 ·
  2. vLLM · Q8_0 (current)Decode 87.69 · Prefill 408.22 ·
  3. Strata · Q3_K_M Decode 208.39 · Prefill 864.96 ·

Core figures

Decode (avg)
87.69 tok/s
Prefill (avg)
408.22 tok/s
TTFT (avg)
0.88 s
VRAM (avg)
1.05 GB
MTP acceptance rate
—
TTFB
0.7 GB
Power draw
91.1 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3-Coder-Next
Quantization
Q8_0
Framework
vLLM
Version
v0.40
Context length
2048 tokens
Batch size
16
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
2x AMD Strix Halo
Measured VRAM (avg)
1.05 GB
OS
Ubuntu 22.04
Driver
amdgpu-pro 24.20
CUDA
N/A
Power draw
91.1 W

Sources & evidence

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

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-16

    v0.40 · Ubuntu 22.04 · CUDA N/A · 2048 ctx

    87.69 tok/s

    Decode

    408.22 tok/s

    Prefill

    0.88 s

    TTFT

    1.05 GB

    VRAM

    —

    MTP

    91.1 W

    Power draw

    L2 batch seed 2026-10-16: vLLM on hw_id=88 running model_id=58 at Q8_0