Arao AI
Back to search results

elephant-1 · Q6_K

Start a discussion
AMD Ryzen AI Max+ 395128GBvLLM v0.25L2 Cross-framework verified

This page aggregates 1 real-world runs of elephant-1 (Q6_K) on AMD Ryzen AI Max+ 395 with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.

22.65 tok/s

Decode

Decode speed

118.27 tok/s

Prefill

Prefill speed

1.87 s

TTFT

Time to first token

21 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

13 days ago

Last verified

Performance

  1. llama.cpp · Q3_K_L Decode 32.72 · Prefill 137.17 ·
  2. vLLM · Q6_K (current)Decode 22.65 · Prefill 118.27 ·
  3. vLLM · Q8_0 Decode 19.62 · Prefill 103.78 ·

Core figures

Decode (avg)
22.65 tok/s
Prefill (avg)
118.27 tok/s
TTFT (avg)
1.87 s
VRAM (avg)
21 GB
MTP acceptance rate
—
TTFB
0.18 GB
Power draw
63.5 W

Configuration

Member-level fields are taken from the most recent run

Model
elephant-1
Quantization
Q6_K
Framework
vLLM
Version
v0.25
Context length
16384 tokens
Batch size
16
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 Ryzen AI Max+ 395
Nominal VRAM
128 GB
Measured VRAM (avg)
21 GB
OS
Ubuntu 24.04
Driver
amdgpu 6.11
CUDA
12.6
Power draw
63.5 W

Get started

Original model resources

Base model resources may not match this quantization. Use the verified deployment weights above when available.

Sources & evidence

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

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-22

    v0.25 · Ubuntu 24.04 · CUDA 12.6 · 16384 ctx

    22.65 tok/s

    Decode

    118.27 tok/s

    Prefill

    1.87 s

    TTFT

    21 GB

    VRAM

    —

    MTP

    63.5 W

    Power draw

    L2 batch seed 2026-10-22: vLLM on hw_id=24 running model_id=95 at Q6_K