Arao AI
Back to search results

Llama-3.1-8B-Instruct · Q3_K_M

Start a discussion
AMD Radeon RX 7900 XTX24GBvLLM v0.21L2 Cross-framework verified

This page aggregates 1 real-world runs of Llama-3.1-8B-Instruct (Q3_K_M) on AMD Radeon RX 7900 XTX with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.

114.54 tok/s

Decode

Decode speed

374.54 tok/s

Prefill

Prefill speed

1.68 s

TTFT

Time to first token

13.52 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 · Q3_K_S Decode 102.17 · Prefill 424.1 ·
  2. llama.cpp · Q4_K_M Decode 83.47 · Prefill 411.3 ·
  3. vLLM · Q3_K_M (current)Decode 114.54 · Prefill 374.54 ·
  4. vLLM · Q4_K_M Decode 87.13 · Prefill 426.78 ·
  5. vLLM · IQ4_XS Decode 91.85 · Prefill 357.41 ·
  6. Strata · IQ4_XS Decode 138.01 · Prefill 527.22 ·

Core figures

Decode (avg)
114.54 tok/s
Prefill (avg)
374.54 tok/s
TTFT (avg)
1.68 s
VRAM (avg)
13.52 GB
MTP acceptance rate
—
TTFB
0.92 GB
Power draw
227.4 W

Configuration

Member-level fields are taken from the most recent run

Model
Llama-3.1-8B-Instruct
Quantization
Q3_K_M
Framework
vLLM
Version
v0.21
Context length
16384 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
AMD Radeon RX 7900 XTX
Nominal VRAM
24 GB
Measured VRAM (avg)
13.52 GB
OS
Windows 11 23H2
Driver
amdgpu-pro 24.20
CUDA
12.8
Power draw
227.4 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

    v0.21 · Windows 11 23H2 · CUDA 12.8 · 16384 ctx

    114.54 tok/s

    Decode

    374.54 tok/s

    Prefill

    1.68 s

    TTFT

    13.52 GB

    VRAM

    —

    MTP

    227.4 W

    Power draw

    L2 batch seed 2026-10-19: vLLM on hw_id=11 running model_id=68 at Q3_K_M