Qwen3-4B-Thinking-2507 · Q3_K_L

Apple M4 Pro64GBStrata v0.25L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3-4B-Thinking-2507 (Q3_K_L) on Apple M4 Pro with Strata, contributed by 1 independent source platforms; metrics are averages of published measurements.

206.28 tok/s

Decode

Decode speed

783.16 tok/s

Prefill

Prefill speed

2.22 s

TTFT

Time to first token

8.15 GB

VRAM

VRAM usage

L2 Cross-framework verified

Measured runs

Independent sources

Other

Source platforms

5 days ago

Last verified

Performance

llama.cpp · IQ4_XS Decode 138.37 · Prefill 661.86 ·
llama.cpp · Q4_K_S Decode 123.54 · Prefill 441.3 ·
Strata · IQ3_XS Decode 218.39 · Prefill 1145.06 ·
Strata · Q3_K_L (current)Decode 206.28 · Prefill 783.16 ·

Core figures

Decode (avg): 206.28 tok/s
Prefill (avg): 783.16 tok/s
TTFT (avg): 2.22 s
VRAM (avg): 8.15 GB
MTP acceptance rate: 0.66
TTFB: 1.65 GB
Power draw: 96 W

Configuration

Member-level fields are taken from the most recent run

Model: Qwen3-4B-Thinking-2507
Quantization: Q3_K_L
Framework: Strata
Version: v0.25
Context length: 32768 tokens
Batch size: 256
GPU layers: 8
Flash Attention: On

Hardware

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

GPU: Apple M4 Pro
Nominal VRAM: 64 GB
Measured VRAM (avg): 8.15 GB
OS: macOS 15.0
Driver: Apple M4 Pro
CUDA: N/A
Power draw: 96 W

Sources & evidence

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

L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-13
v0.25 · macOS 15.0 · CUDA N/A · 32768 ctx
206.28 tok/s
Decode
783.16 tok/s
Prefill
2.22 s
TTFT
8.15 GB
VRAM
0.66
MTP
96 W
Power draw
L2 batch seed 2026-10-13: Strata on hw_id=25 running model_id=57 at Q3_K_L