Arao AI
Back to search results

Hy-Memory-30B-A3B · Q2_K

Start a discussion
Apple M2 Ultrallama.cpp v0.49L2 Cross-framework verified

This page aggregates 1 real-world runs of Hy-Memory-30B-A3B (Q2_K) on Apple M2 Ultra with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.

207.84 tok/s

Decode

Decode speed

1,139.16 tok/s

Prefill

Prefill speed

0.49 s

TTFT

Time to first token

36.75 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

17 days ago

Last verified

Performance

  1. llama.cpp · IQ2_XS Decode 240.85 · Prefill 1025.78 ·
  2. llama.cpp · Q2_K (current)Decode 207.84 · Prefill 1139.16 ·
  3. Ollama · Q2_K Decode 187.09 · Prefill 853.49 ·
  4. Ollama · Q4_K_M Decode 134.47 · Prefill 706.31 ·

Core figures

Decode (avg)
207.84 tok/s
Prefill (avg)
1,139.16 tok/s
TTFT (avg)
0.49 s
VRAM (avg)
36.75 GB
MTP acceptance rate
—
TTFB
1.66 GB
Power draw
149.8 W

Configuration

Member-level fields are taken from the most recent run

Model
Hy-Memory-30B-A3B
Quantization
Q2_K
Framework
llama.cpp
Version
v0.49
Context length
2048 tokens
Batch size
256
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)
36.75 GB
OS
macOS 26.1
Power draw
149.8 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 2026-10-19

    v0.49 · macOS 26.1 · 2048 ctx

    207.84 tok/s

    Decode

    1,139.16 tok/s

    Prefill

    0.49 s

    TTFT

    36.75 GB

    VRAM

    —

    MTP

    149.8 W

    Power draw

    L2 batch seed 2026-10-19: llama.cpp on hw_id=103 running model_id=87 at Q2_K