Arao AI
Back to search results

Qwen3.5-0.8B · Q5_K_S

Start a discussion
NVIDIA RTX 309024GBllama.cpp v0.50L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3.5-0.8B (Q5_K_S) on NVIDIA RTX 3090 with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.

157.88 tok/s

Decode

Decode speed

604.44 tok/s

Prefill

Prefill speed

0.39 s

TTFT

Time to first token

3.61 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 · Q5_K_S (current)Decode 157.88 · Prefill 604.44 ·
  2. Ollama · IQ2_M Decode 246.2 · Prefill 1112.28 ·
  3. vLLM · Q3_K_M Decode 217.39 · Prefill 922.41 ·
  4. vLLM · Q3_K_S Decode 241.45 · Prefill 1059.79 ·
  5. Ollama · Q3_K_L Decode 178.6 · Prefill 868.06 ·
  6. Ollama · Q5_K_M Decode 146.73 · Prefill 547.3 ·
  7. llama.cpp · IQ2_M Decode 352.68 · Prefill 1697.91 ·
  8. llama.cpp · Q3_K_L Decode 315.76 · Prefill 1365.7 ·
  9. Strata · IQ4_XS Decode 382.98 · Prefill 1282.3 ·
  10. Strata · Q3_K_M Decode 336.19 · Prefill 1574.41 ·

Core figures

Decode (avg)
157.88 tok/s
Prefill (avg)
604.44 tok/s
TTFT (avg)
0.39 s
VRAM (avg)
3.61 GB
MTP acceptance rate
—
TTFB
1.26 GB
Power draw
297.7 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3.5-0.8B
Quantization
Q5_K_S
Framework
llama.cpp
Version
v0.50
Context length
16384 tokens
Batch size
512
GPU layers
99
Flash Attention
Off

Hardware

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

GPU
NVIDIA RTX 3090
Nominal VRAM
24 GB
Measured VRAM (avg)
3.61 GB
OS
Ubuntu 22.04
Driver
565.57
CUDA
12.6
Power draw
297.7 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

    v0.50 · Ubuntu 22.04 · CUDA 12.6 · 16384 ctx

    157.88 tok/s

    Decode

    604.44 tok/s

    Prefill

    0.39 s

    TTFT

    3.61 GB

    VRAM

    —

    MTP

    297.7 W

    Power draw

    L2 batch seed 2026-10-19: llama.cpp on hw_id=2 running model_id=51 at Q5_K_S