Arao AI
Back to search results

Qwen3-4B · Q3_K_L

Start a discussion
NVIDIA RTX 4070 Ti12GBllama.cpp v0.23L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3-4B (Q3_K_L) on NVIDIA RTX 4070 Ti with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.

107.42 tok/s

Decode

Decode speed

453.38 tok/s

Prefill

Prefill speed

1.96 s

TTFT

Time to first token

8.15 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

5 days ago

Last verified

Performance

  1. llama.cpp · Q3_K_L (current)Decode 107.42 · Prefill 453.38 ·
  2. llama.cpp · Q4_K_M Decode 104 · Prefill 443.35 ·
  3. vLLM · Q8_0 Decode 66.34 · Prefill 241.62 ·

Core figures

Decode (avg)
107.42 tok/s
Prefill (avg)
453.38 tok/s
TTFT (avg)
1.96 s
VRAM (avg)
8.15 GB
MTP acceptance rate
—
TTFB
0.86 GB
Power draw
275.4 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3-4B
Quantization
Q3_K_L
Framework
llama.cpp
Version
v0.23
Context length
32768 tokens
Batch size
512
GPU layers
99
Flash Attention
On

Hardware

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

GPU
NVIDIA RTX 4070 Ti
Nominal VRAM
12 GB
Measured VRAM (avg)
8.15 GB
OS
Ubuntu 24.04
Driver
565.57
CUDA
12.4
Power draw
275.4 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

    v0.23 · Ubuntu 24.04 · CUDA 12.4 · 32768 ctx

    107.42 tok/s

    Decode

    453.38 tok/s

    Prefill

    1.96 s

    TTFT

    8.15 GB

    VRAM

    —

    MTP

    275.4 W

    Power draw

    L2 batch seed 2026-10-13: llama.cpp on hw_id=42 running model_id=54 at Q3_K_L