Arao AI
Back to search results

gemma-4-12B-it · IQ2_XS

Start a discussion
NVIDIA RTX 508016GBllama.cpp v0.42L2 Cross-framework verified

This page aggregates 1 real-world runs of gemma-4-12B-it (IQ2_XS) on NVIDIA RTX 5080 with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.

194.87 tok/s

Decode

Decode speed

964.07 tok/s

Prefill

Prefill speed

1.95 s

TTFT

Time to first token

14.2 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 Decode 201.04 · Prefill 1060.13 ·
  2. llama.cpp · IQ2_XS (current)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)
194.87 tok/s
Prefill (avg)
964.07 tok/s
TTFT (avg)
1.95 s
VRAM (avg)
14.2 GB
MTP acceptance rate
—
TTFB
1.56 GB
Power draw
253.4 W

Configuration

Member-level fields are taken from the most recent run

Model
gemma-4-12B-it
Quantization
IQ2_XS
Framework
llama.cpp
Version
v0.42
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
NVIDIA RTX 5080
Nominal VRAM
16 GB
Measured VRAM (avg)
14.2 GB
OS
Ubuntu 22.04
Driver
560.94
CUDA
12.6
Power draw
253.4 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

    v0.42 · Ubuntu 22.04 · CUDA 12.6 · 4096 ctx

    194.87 tok/s

    Decode

    964.07 tok/s

    Prefill

    1.95 s

    TTFT

    14.2 GB

    VRAM

    —

    MTP

    253.4 W

    Power draw

    L2 batch seed 2026-10-18: llama.cpp on hw_id=14 running model_id=65 at IQ2_XS