Arao AI
Back to search results

Qwen3.5-4B · Q3_K_L

Start a discussion
NVIDIA RTX 4070 Ti12GBllama.cpp v0.49L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3.5-4B (Q3_K_L) on NVIDIA RTX 4070 Ti with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.

123.86 tok/s

Decode

Decode speed

681.14 tok/s

Prefill

Prefill speed

1.87 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. vLLM · Q4_K_S Decode 161.78 · Prefill 577.24 ·
  2. Ollama · Q6_K Decode 99.87 · Prefill 394.75 ·
  3. llama.cpp · Q3_K_L (current)Decode 123.86 · Prefill 681.14 ·
  4. Ollama · Q8_0 Decode 68.74 · Prefill 373.18 ·

Core figures

Decode (avg)
123.86 tok/s
Prefill (avg)
681.14 tok/s
TTFT (avg)
1.87 s
VRAM (avg)
8.15 GB
MTP acceptance rate
—
TTFB
0.99 GB
Power draw
219.4 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3.5-4B
Quantization
Q3_K_L
Framework
llama.cpp
Version
v0.49
Context length
4096 tokens
Batch size
256
GPU layers
33
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 22.04
Driver
570.86
CUDA
12.8
Power draw
219.4 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

    v0.49 · Ubuntu 22.04 · CUDA 12.8 · 4096 ctx

    123.86 tok/s

    Decode

    681.14 tok/s

    Prefill

    1.87 s

    TTFT

    8.15 GB

    VRAM

    —

    MTP

    219.4 W

    Power draw

    L2 batch seed 2026-10-13: llama.cpp on hw_id=42 running model_id=48 at Q3_K_L