Arao AI
Back to search results

Qwen2.5-Omni-3B · Q3_K_M

Start a discussion
Qualcomm Snapdragon 7 Gen 4llama.cpp v0.46L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen2.5-Omni-3B (Q3_K_M) on Qualcomm Snapdragon 7 Gen 4 with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.

59.3 tok/s

Decode

Decode speed

285.77 tok/s

Prefill

Prefill speed

0.37 s

TTFT

Time to first token

6.38 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

2 days ago

Last verified

Performance

  1. llama.cpp · Q3_K_M (current)Decode 59.3 · Prefill 285.77 ·
  2. Strata · Q3_K_S Decode 97.85 · Prefill 474.54 ·
  3. Strata · Q5_K_M Decode 64.91 · Prefill 252.46 ·

Core figures

Decode (avg)
59.3 tok/s
Prefill (avg)
285.77 tok/s
TTFT (avg)
0.37 s
VRAM (avg)
6.38 GB
MTP acceptance rate
—
TTFB
0.47 GB
Power draw
25.6 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen2.5-Omni-3B
Quantization
Q3_K_M
Framework
llama.cpp
Version
v0.46
Context length
2048 tokens
Batch size
512
GPU layers
20
Flash Attention
Off

Hardware

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

GPU
Qualcomm Snapdragon 7 Gen 4
Measured VRAM (avg)
6.38 GB
OS
Windows 11 23H2
Driver
Qualcomm Adreno 7.0
CUDA
N/A
Power draw
25.6 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

    v0.46 · Windows 11 23H2 · CUDA N/A · 2048 ctx

    59.3 tok/s

    Decode

    285.77 tok/s

    Prefill

    0.37 s

    TTFT

    6.38 GB

    VRAM

    —

    MTP

    25.6 W

    Power draw

    L2 batch seed 2026-10-16: llama.cpp on hw_id=105 running model_id=60 at Q3_K_M