Arao AI
Back to search results

gemma-4-26B-A4B-it · IQ2_XS

Start a discussion
NVIDIA RTX 5080 SUPER24GBllama.cpp b4200L0 Self-reported

This page aggregates 1 real-world runs of gemma-4-26B-A4B-it (IQ2_XS) on NVIDIA RTX 5080 SUPER with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.

99.3 tok/s

Decode

Decode speed

744.8 tok/s

Prefill

Prefill speed

0.13 s

TTFT

Time to first token

9.3 GB

VRAM

VRAM usage

L0 Self-reported

1

Measured runs

1

Independent sources

Other

Source platforms

Today

Last verified

Performance

  1. vLLM · IQ2_XS Decode 108.2 · Prefill 811.5 ·
  2. vLLM · Q5_K_M Decode 49.3 · Prefill 369.8 ·
  3. vLLM · Q4_K_M Decode 60.3 · Prefill 452.3 ·
  4. llama.cpp · IQ2_XS (current)Decode 99.3 · Prefill 744.8 ·
  5. llama.cpp · Q5_K_M Decode 45.3 · Prefill 339.8 ·
  6. llama.cpp · Q4_K_M Decode 55.3 · Prefill 414.8 ·

Core figures

Decode (avg)
99.3 tok/s
Prefill (avg)
744.8 tok/s
TTFT (avg)
0.13 s
VRAM (avg)
9.3 GB
MTP acceptance rate
—
TTFB
— GB
Power draw
— W

Configuration

Member-level fields are taken from the most recent run

Model
gemma-4-26B-A4B-it
Quantization
IQ2_XS
Framework
llama.cpp
Version
b4200
Context length
4096 tokens
Batch size
512
GPU layers
99
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 SUPER
Nominal VRAM
24 GB
Measured VRAM (avg)
9.3 GB
OS
Linux
Driver
570.65
CUDA
12.6

Sources & evidence

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

  1. L0 Self-reportedOtherOriginal link Verified on 2026-10-06

    b4200 · Linux · CUDA 12.6 · 4096 ctx

    99.3 tok/s

    Decode

    744.8 tok/s

    Prefill

    0.13 s

    TTFT

    9.3 GB

    VRAM

    —

    MTP

    — W

    Power draw

    L0 理论估算（add-1000-records-batch3）