Arao AI
Back to search results

mammoth-1 · IQ1_S

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

This page aggregates 1 real-world runs of mammoth-1 (IQ1_S) on Apple M2 Ultra with Ollama, contributed by 1 independent source platforms; metrics are averages of published measurements.

100.6 tok/s

Decode

Decode speed

448.02 tok/s

Prefill

Prefill speed

2.5 s

TTFT

Time to first token

25.91 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 · IQ2_M Decode 102.29 · Prefill 417.96 ·
  2. llama.cpp · Q5_K_M Decode 67.37 · Prefill 314.26 ·
  3. Ollama · IQ1_S (current)Decode 100.6 · Prefill 448.02 ·

Core figures

Decode (avg)
100.6 tok/s
Prefill (avg)
448.02 tok/s
TTFT (avg)
2.5 s
VRAM (avg)
25.91 GB
MTP acceptance rate
—
TTFB
0.8 GB
Power draw
111 W

Configuration

Member-level fields are taken from the most recent run

Model
mammoth-1
Quantization
IQ1_S
Framework
Ollama
Version
v0.52
Context length
2048 tokens
Batch size
512
GPU layers
20
Flash Attention
Off

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)
25.91 GB
OS
macOS 26.1
Power draw
111 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

    v0.52 · macOS 26.1 · 2048 ctx

    100.6 tok/s

    Decode

    448.02 tok/s

    Prefill

    2.5 s

    TTFT

    25.91 GB

    VRAM

    —

    MTP

    111 W

    Power draw

    L2 batch seed 2026-10-22: Ollama on hw_id=103 running model_id=94 at IQ1_S