Arao AI
Back to search results

gemma-4-12B-it · IQ2_M

Start a discussion
AMD Ryzen AI Max+ 395128GBllama.cpp v0.24L2 Cross-framework verified

This page aggregates 1 real-world runs of gemma-4-12B-it (IQ2_M) on AMD Ryzen AI Max+ 395 with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.

32.57 tok/s

Decode

Decode speed

131.77 tok/s

Prefill

Prefill speed

2.15 s

TTFT

Time to first token

15.2 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_M (current)Decode 32.57 · Prefill 131.77 ·
  2. llama.cpp · Q5_K_M Decode 22.71 · Prefill 109.92 ·
  3. Ollama · Q3_K_M Decode 26.82 · Prefill 137.08 ·
  4. Ollama · Q5_K_S Decode 20.2 · Prefill 85.22 ·
  5. Ollama · Q4_K_M Decode 66.99 · Prefill 227.06 ·
  6. Strata · Q8_0 Decode 71.19 · Prefill 287.95 ·

Core figures

Decode (avg)
32.57 tok/s
Prefill (avg)
131.77 tok/s
TTFT (avg)
2.15 s
VRAM (avg)
15.2 GB
MTP acceptance rate
—
TTFB
0.26 GB
Power draw
62 W

Configuration

Member-level fields are taken from the most recent run

Model
gemma-4-12B-it
Quantization
IQ2_M
Framework
llama.cpp
Version
v0.24
Context length
4096 tokens
Batch size
512
GPU layers
33
Flash Attention
On

Hardware

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

GPU
AMD Ryzen AI Max+ 395
Nominal VRAM
128 GB
Measured VRAM (avg)
15.2 GB
OS
Windows 11 23H2
Driver
amdgpu 6.12
CUDA
12.4
Power draw
62 W

Sources & evidence

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

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-19

    v0.24 · Windows 11 23H2 · CUDA 12.4 · 4096 ctx

    32.57 tok/s

    Decode

    131.77 tok/s

    Prefill

    2.15 s

    TTFT

    15.2 GB

    VRAM

    —

    MTP

    62 W

    Power draw

    L2 batch seed 2026-10-19: llama.cpp on hw_id=24 running model_id=65 at IQ2_M