Arao AI
Back to search results

Qwen3.6-35B-A3B · IQ3_XS

Start a discussion
NVIDIA RTX PRO 6000 Blackwell 96GB96GBvLLM v0.23L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3.6-35B-A3B (IQ3_XS) on NVIDIA RTX PRO 6000 Blackwell 96GB with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.

319.33 tok/s

Decode

Decode speed

1,553.22 tok/s

Prefill

Prefill speed

0.32 s

TTFT

Time to first token

47.67 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

17 days ago

Last verified

Performance

  1. vLLM · IQ3_XS (current)Decode 319.33 · Prefill 1553.22 ·
  2. Ollama · Q5_K_S Decode 301.35 · Prefill 1440.97 ·

Core figures

Decode (avg)
319.33 tok/s
Prefill (avg)
1,553.22 tok/s
TTFT (avg)
0.32 s
VRAM (avg)
47.67 GB
MTP acceptance rate
—
TTFB
2.55 GB
Power draw
331.5 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3.6-35B-A3B
Quantization
IQ3_XS
Framework
vLLM
Version
v0.23
Context length
4096 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 PRO 6000 Blackwell 96GB
Nominal VRAM
96 GB
Measured VRAM (avg)
47.67 GB
OS
Windows 11 23H2
Driver
560.94
CUDA
12.4
Power draw
331.5 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

    v0.23 · Windows 11 23H2 · CUDA 12.4 · 4096 ctx

    319.33 tok/s

    Decode

    1,553.22 tok/s

    Prefill

    0.32 s

    TTFT

    47.67 GB

    VRAM

    —

    MTP

    331.5 W

    Power draw

    L2 batch seed 2026-10-19: vLLM on hw_id=68 running model_id=47 at IQ3_XS