Arao AI
Back to search results

gemma-4-12B-it · IQ2_M

Start a discussion
NVIDIA RTX 508016GBllama.cpp v0.28L2 Cross-framework verified

This page aggregates 1 real-world runs of gemma-4-12B-it (IQ2_M) on NVIDIA RTX 5080 with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.

201.04 tok/s

Decode

Decode speed

1,060.13 tok/s

Prefill

Prefill speed

0.38 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

Today

Last verified

Performance

  1. llama.cpp · IQ2_M (current)Decode 201.04 · Prefill 1060.13 ·
  2. llama.cpp · IQ2_XS Decode 194.87 · Prefill 964.07 ·
  3. Strata · IQ2_M Decode 277.55 · Prefill 1220.39 ·
  4. Ollama · IQ2_XS Decode 103.79 · Prefill 371.24 ·

Core figures

Decode (avg)
201.04 tok/s
Prefill (avg)
1,060.13 tok/s
TTFT (avg)
0.38 s
VRAM (avg)
15.2 GB
MTP acceptance rate
—
TTFB
1.61 GB
Power draw
244.1 W

Configuration

Member-level fields are taken from the most recent run

Model
gemma-4-12B-it
Quantization
IQ2_M
Framework
llama.cpp
Version
v0.28
Context length
2048 tokens
Batch size
256
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
NVIDIA RTX 5080
Nominal VRAM
16 GB
Measured VRAM (avg)
15.2 GB
OS
Windows 11 23H2
Driver
560.94
CUDA
12.8
Power draw
244.1 W

Sources & evidence

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

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-18

    v0.28 · Windows 11 23H2 · CUDA 12.8 · 2048 ctx

    201.04 tok/s

    Decode

    1,060.13 tok/s

    Prefill

    0.38 s

    TTFT

    15.2 GB

    VRAM

    —

    MTP

    244.1 W

    Power draw

    L2 batch seed 2026-10-18: llama.cpp on hw_id=14 running model_id=65 at IQ2_M