Arao AI
Back to search results

Qwen3-4B-Thinking-2507 · Q3_K_L

Start a discussion
NVIDIA RTX 4060 Ti 16GB16GBStrata v0.21L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3-4B-Thinking-2507 (Q3_K_L) on NVIDIA RTX 4060 Ti 16GB with Strata, contributed by 1 independent source platforms; metrics are averages of published measurements.

140.78 tok/s

Decode

Decode speed

483.92 tok/s

Prefill

Prefill speed

0.93 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 · IQ3_S Decode 108.19 · Prefill 350.2 ·
  2. vLLM · Q3_K_M Decode 119.54 · Prefill 394.34 ·
  3. Strata · IQ3_XS Decode 160.63 · Prefill 538.1 ·
  4. Strata · Q3_K_L (current)Decode 140.78 · Prefill 483.92 ·

Core figures

Decode (avg)
140.78 tok/s
Prefill (avg)
483.92 tok/s
TTFT (avg)
0.93 s
VRAM (avg)
8.15 GB
MTP acceptance rate
0.63
TTFB
1.13 GB
Power draw
310.2 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3-4B-Thinking-2507
Quantization
Q3_K_L
Framework
Strata
Version
v0.21
Context length
16384 tokens
Batch size
1024
GPU layers
5
Flash Attention
On

Hardware

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

GPU
NVIDIA RTX 4060 Ti 16GB
Nominal VRAM
16 GB
Measured VRAM (avg)
8.15 GB
OS
macOS 15.0
Driver
565.57
CUDA
12.8
Power draw
310.2 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 12.8 · 16384 ctx

    140.78 tok/s

    Decode

    483.92 tok/s

    Prefill

    0.93 s

    TTFT

    8.15 GB

    VRAM

    0.63

    MTP

    310.2 W

    Power draw

    L2 batch seed 2026-10-13: Strata on hw_id=3 running model_id=57 at Q3_K_L