Arao AI
Back to search results

elephant-1 · Q4_K_S

Start a discussion
Apple M2 Ultrallama.cpp v0.35L2 Cross-framework verified

This page aggregates 1 real-world runs of elephant-1 (Q4_K_S) on Apple M2 Ultra with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.

88.6 tok/s

Decode

Decode speed

484.39 tok/s

Prefill

Prefill speed

1.14 s

TTFT

Time to first token

16.84 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

13 days ago

Last verified

Performance

  1. llama.cpp · Q4_K_M Decode 92.52 · Prefill 377.54 ·
  2. llama.cpp · Q4_K_S (current)Decode 88.6 · Prefill 484.39 ·
  3. Ollama · Q6_K Decode 67.84 · Prefill 354.32 ·

Core figures

Decode (avg)
88.6 tok/s
Prefill (avg)
484.39 tok/s
TTFT (avg)
1.14 s
VRAM (avg)
16.84 GB
MTP acceptance rate
—
TTFB
0.71 GB
Power draw
121.5 W

Configuration

Member-level fields are taken from the most recent run

Model
elephant-1
Quantization
Q4_K_S
Framework
llama.cpp
Version
v0.35
Context length
4096 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)
16.84 GB
OS
macOS 26.1
Power draw
121.5 W

Get started

Original model resources

Base model resources may not match this quantization. Use the verified deployment weights above when available.

Sources & evidence

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

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-22

    v0.35 · macOS 26.1 · 4096 ctx

    88.6 tok/s

    Decode

    484.39 tok/s

    Prefill

    1.14 s

    TTFT

    16.84 GB

    VRAM

    —

    MTP

    121.5 W

    Power draw

    L2 batch seed 2026-10-22: llama.cpp on hw_id=103 running model_id=95 at Q4_K_S