Arao AI
Back to search results

Ministral-3-8B-Instruct-2512 · Q3_K_M

Start a discussion
Apple M2 Ultrallama.cpp v0.52L2 Cross-framework verified

This page aggregates 1 real-world runs of Ministral-3-8B-Instruct-2512 (Q3_K_M) on Apple M2 Ultra with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.

119.5 tok/s

Decode

Decode speed

580.12 tok/s

Prefill

Prefill speed

1.82 s

TTFT

Time to first token

13.52 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

11 days ago

Last verified

Performance

  1. llama.cpp · Q3_K_M (current)Decode 119.5 · Prefill 580.12 ·
  2. Ollama · IQ3_S Decode 108.26 · Prefill 586.93 ·

Core figures

Decode (avg)
119.5 tok/s
Prefill (avg)
580.12 tok/s
TTFT (avg)
1.82 s
VRAM (avg)
13.52 GB
MTP acceptance rate
—
TTFB
0.96 GB
Power draw
148.6 W

Configuration

Member-level fields are taken from the most recent run

Model
Ministral-3-8B-Instruct-2512
Quantization
Q3_K_M
Framework
llama.cpp
Version
v0.52
Context length
65536 tokens
Batch size
512
GPU layers
28
Flash Attention
On

Hardware

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

GPU
Apple M2 Ultra
Measured VRAM (avg)
13.52 GB
OS
macOS 26.1
Power draw
148.6 W

Sources & evidence

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

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-24

    v0.52 · macOS 26.1 · 65536 ctx

    119.5 tok/s

    Decode

    580.12 tok/s

    Prefill

    1.82 s

    TTFT

    13.52 GB

    VRAM

    —

    MTP

    148.6 W

    Power draw

    L2 batch seed 2026-10-24: llama.cpp on hw_id=103 running model_id=74 at Q3_K_M