Arao AI
Back to search results

Qwen3-4B-Thinking-2507 · Q5_K_M

Start a discussion
NVIDIA RTX 408016GBvLLM v0.21L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3-4B-Thinking-2507 (Q5_K_M) on NVIDIA RTX 4080 with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.

92.54 tok/s

Decode

Decode speed

405.16 tok/s

Prefill

Prefill speed

1.01 s

TTFT

Time to first token

9.91 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

5 days ago

Last verified

Performance

  1. vLLM · Q5_K_M (current)Decode 92.54 · Prefill 405.16 ·

Core figures

Decode (avg)
92.54 tok/s
Prefill (avg)
405.16 tok/s
TTFT (avg)
1.01 s
VRAM (avg)
9.91 GB
MTP acceptance rate
—
TTFB
0.74 GB
Power draw
366.3 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3-4B-Thinking-2507
Quantization
Q5_K_M
Framework
vLLM
Version
v0.21
Context length
16384 tokens
Batch size
16
GPU layers
999
Flash Attention
Off

Hardware

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

GPU
NVIDIA RTX 4080
Nominal VRAM
16 GB
Measured VRAM (avg)
9.91 GB
OS
macOS 15.0
Driver
570.86
CUDA
12.6
Power draw
366.3 W

Sources & evidence

1 measured records in total, each traceable to its original source

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-13

    v0.21 · macOS 15.0 · CUDA 12.6 · 16384 ctx

    92.54 tok/s

    Decode

    405.16 tok/s

    Prefill

    1.01 s

    TTFT

    9.91 GB

    VRAM

    —

    MTP

    366.3 W

    Power draw

    L2 batch seed 2026-10-13: vLLM on hw_id=41 running model_id=57 at Q5_K_M