Arao AI
Back to search results

DeepSeek-R1-Distill-Qwen-14B · IQ1_S

Start a discussion
Apple M5 Max128GBllama.cpp v0.39L2 Cross-framework verified

This page aggregates 1 real-world runs of DeepSeek-R1-Distill-Qwen-14B (IQ1_S) on Apple M5 Max with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.

202.22 tok/s

Decode

Decode speed

672.97 tok/s

Prefill

Prefill speed

1.5 s

TTFT

Time to first token

14.45 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 · IQ1_S (current)Decode 202.22 · Prefill 672.97 ·
  2. vLLM · Q4_K_M Decode 134.81 · Prefill 469.71 ·

Core figures

Decode (avg)
202.22 tok/s
Prefill (avg)
672.97 tok/s
TTFT (avg)
1.5 s
VRAM (avg)
14.45 GB
MTP acceptance rate
—
TTFB
1.62 GB
Power draw
131.8 W

Configuration

Member-level fields are taken from the most recent run

Model
DeepSeek-R1-Distill-Qwen-14B
Quantization
IQ1_S
Framework
llama.cpp
Version
v0.39
Context length
2048 tokens
Batch size
256
GPU layers
99
Flash Attention
Off

Hardware

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

GPU
Apple M5 Max
Nominal VRAM
128 GB
Measured VRAM (avg)
14.45 GB
OS
macOS 15.0
Driver
Apple M5 Max
CUDA
N/A
Power draw
131.8 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

    v0.39 · macOS 15.0 · CUDA N/A · 2048 ctx

    202.22 tok/s

    Decode

    672.97 tok/s

    Prefill

    1.5 s

    TTFT

    14.45 GB

    VRAM

    —

    MTP

    131.8 W

    Power draw

    L2 batch seed 2026-10-13: llama.cpp on hw_id=28 running model_id=71 at IQ1_S