Arao AI
Back to search results

Qwen3-4B · Q2_K

Start a discussion
NVIDIA Jetson AGX Orin 64GB64GBvLLM v0.26L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3-4B (Q2_K) on NVIDIA Jetson AGX Orin 64GB with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.

95.78 tok/s

Decode

Decode speed

454.64 tok/s

Prefill

Prefill speed

2.45 s

TTFT

Time to first token

6.72 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 · IQ2_M Decode 92.76 · Prefill 306.43 ·
  2. vLLM · IQ2_M Decode 96.3 · Prefill 512.46 ·
  3. vLLM · Q2_K (current)Decode 95.78 · Prefill 454.64 ·
  4. Ollama · IQ3_S Decode 73.25 · Prefill 282.68 ·
  5. Ollama · Q4_K_M Decode 58.47 · Prefill 300.76 ·

Core figures

Decode (avg)
95.78 tok/s
Prefill (avg)
454.64 tok/s
TTFT (avg)
2.45 s
VRAM (avg)
6.72 GB
MTP acceptance rate
—
TTFB
0.77 GB
Power draw
44.1 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3-4B
Quantization
Q2_K
Framework
vLLM
Version
v0.26
Context length
2048 tokens
Batch size
8
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 Jetson AGX Orin 64GB
Nominal VRAM
64 GB
Measured VRAM (avg)
6.72 GB
OS
JetPack 6.0
Driver
NVIDIA L4T 36.3
CUDA
12.6
Power draw
44.1 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

    v0.26 · JetPack 6.0 · CUDA 12.6 · 2048 ctx

    95.78 tok/s

    Decode

    454.64 tok/s

    Prefill

    2.45 s

    TTFT

    6.72 GB

    VRAM

    —

    MTP

    44.1 W

    Power draw

    L2 batch seed 2026-10-16: vLLM on hw_id=76 running model_id=54 at Q2_K