Arao AI
Back to search results

Qwen3.5-4B · Q3_K_M

Start a discussion
NVIDIA RTX PRO 6000 Blackwell 96GB96GBvLLM v0.39L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3.5-4B (Q3_K_M) on NVIDIA RTX PRO 6000 Blackwell 96GB with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.

336.72 tok/s

Decode

Decode speed

1,847.35 tok/s

Prefill

Prefill speed

2.46 s

TTFT

Time to first token

7.81 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

16 days ago

Last verified

Performance

  1. vLLM · Q3_K_M (current)Decode 336.72 · Prefill 1847.35 ·
  2. vLLM · Q6_K Decode 245.04 · Prefill 1242.69 ·
  3. Ollama · Q3_K_L Decode 279.88 · Prefill 1395.76 ·
  4. Ollama · Q5_K_S Decode 216.95 · Prefill 820.67 ·

Core figures

Decode (avg)
336.72 tok/s
Prefill (avg)
1,847.35 tok/s
TTFT (avg)
2.46 s
VRAM (avg)
7.81 GB
MTP acceptance rate
—
TTFB
2.69 GB
Power draw
385.9 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3.5-4B
Quantization
Q3_K_M
Framework
vLLM
Version
v0.39
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 RTX PRO 6000 Blackwell 96GB
Nominal VRAM
96 GB
Measured VRAM (avg)
7.81 GB
OS
Windows 11 24H2
Driver
560.94
CUDA
12.8
Power draw
385.9 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

    v0.39 · Windows 11 24H2 · CUDA 12.8 · 2048 ctx

    336.72 tok/s

    Decode

    1,847.35 tok/s

    Prefill

    2.46 s

    TTFT

    7.81 GB

    VRAM

    —

    MTP

    385.9 W

    Power draw

    L2 batch seed 2026-10-19: vLLM on hw_id=68 running model_id=48 at Q3_K_M