Arao AI
Back to search results

gemma-4-12B-it · Q3_K_M

Start a discussion
AMD Radeon RX 7900 XTX24GBvLLM v0.48L2 Cross-framework verified

This page aggregates 1 real-world runs of gemma-4-12B-it (Q3_K_M) on AMD Radeon RX 7900 XTX with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.

77.87 tok/s

Decode

Decode speed

405.9 tok/s

Prefill

Prefill speed

0.18 s

TTFT

Time to first token

19.24 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

17 days ago

Last verified

Performance

  1. vLLM · Q3_K_M (current)Decode 77.87 · Prefill 405.9 ·
  2. Ollama · IQ1_S Decode 92.04 · Prefill 422.92 ·
  3. llama.cpp · IQ2_XS Decode 97.1 · Prefill 483.73 ·
  4. llama.cpp · Q5_K_M Decode 54.12 · Prefill 201.79 ·
  5. Strata · IQ2_M Decode 125.31 · Prefill 624.12 ·
  6. vLLM · Q4_K_S Decode 86.73 · Prefill 462.81 ·
  7. vLLM · Q5_K_S Decode 64.98 · Prefill 291.13 ·
  8. Ollama · IQ2_M Decode 92.49 · Prefill 473.64 ·
  9. Ollama · Q4_K_S Decode 67.64 · Prefill 313.4 ·
  10. Strata · IQ3_XS Decode 129.13 · Prefill 417.63 ·

Core figures

Decode (avg)
77.87 tok/s
Prefill (avg)
405.9 tok/s
TTFT (avg)
0.18 s
VRAM (avg)
19.24 GB
MTP acceptance rate
—
TTFB
0.62 GB
Power draw
236 W

Configuration

Member-level fields are taken from the most recent run

Model
gemma-4-12B-it
Quantization
Q3_K_M
Framework
vLLM
Version
v0.48
Context length
32768 tokens
Batch size
16
GPU layers
999
Flash Attention
On

Hardware

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

GPU
AMD Radeon RX 7900 XTX
Nominal VRAM
24 GB
Measured VRAM (avg)
19.24 GB
OS
Ubuntu 24.04
Driver
amdgpu-pro 24.20
CUDA
12.6
Power draw
236 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

    v0.48 · Ubuntu 24.04 · CUDA 12.6 · 32768 ctx

    77.87 tok/s

    Decode

    405.9 tok/s

    Prefill

    0.18 s

    TTFT

    19.24 GB

    VRAM

    —

    MTP

    236 W

    Power draw

    L2 batch seed 2026-10-19: vLLM on hw_id=11 running model_id=65 at Q3_K_M