Arao AI
Back to search results

Qwen3-Embedding-0.6B · Q3_K_S

Start a discussion
NVIDIA RTX 509032GBvLLM v0.26L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3-Embedding-0.6B (Q3_K_S) on NVIDIA RTX 5090 with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.

322.2 tok/s

Decode

Decode speed

1,743.15 tok/s

Prefill

Prefill speed

2.04 s

TTFT

Time to first token

2.92 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 · IQ2_XS Decode 304.51 · Prefill 1432.77 ·
  2. vLLM · Q3_K_S (current)Decode 322.2 · Prefill 1743.15 ·
  3. Strata · Q2_K Decode 350.56 · Prefill 1533.82 ·
  4. Strata · Q5_K_M Decode 291.92 · Prefill 1133.9 ·

Core figures

Decode (avg)
322.2 tok/s
Prefill (avg)
1,743.15 tok/s
TTFT (avg)
2.04 s
VRAM (avg)
2.92 GB
MTP acceptance rate
—
TTFB
2.58 GB
Power draw
335.6 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3-Embedding-0.6B
Quantization
Q3_K_S
Framework
vLLM
Version
v0.26
Context length
8192 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
NVIDIA RTX 5090
Nominal VRAM
32 GB
Measured VRAM (avg)
2.92 GB
OS
Ubuntu 22.04
Driver
565.57
CUDA
12.4
Power draw
335.6 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

    v0.26 · Ubuntu 22.04 · CUDA 12.4 · 8192 ctx

    322.2 tok/s

    Decode

    1,743.15 tok/s

    Prefill

    2.04 s

    TTFT

    2.92 GB

    VRAM

    —

    MTP

    335.6 W

    Power draw

    L2 batch seed 2026-10-16: vLLM on hw_id=13 running model_id=62 at Q3_K_S