Arao AI
Back to search results

Qwen3.5-4B · Q8_0

Start a discussion
NVIDIA RTX 508016GBStrata v0.37L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3.5-4B (Q8_0) on NVIDIA RTX 5080 with Strata, contributed by 1 independent source platforms; metrics are averages of published measurements.

218.51 tok/s

Decode

Decode speed

985.54 tok/s

Prefill

Prefill speed

0.97 s

TTFT

Time to first token

11.76 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

16 days ago

Last verified

Performance

  1. Ollama · FP16 Decode 104.67 · Prefill 436.92 ·
  2. Ollama · Q4_K_S Decode 194.55 · Prefill 763.79 ·
  3. Strata · Q8_0 (current)Decode 218.51 · Prefill 985.54 ·
  4. llama.cpp · IQ1_S Decode 350.61 · Prefill 1829.73 ·
  5. llama.cpp · IQ2_XS Decode 312.6 · Prefill 1217.96 ·
  6. Strata · IQ3_S Decode 356.3 · Prefill 1901.07 ·
  7. llama.cpp · IQ4_XS Decode 133.14 · Prefill 731.5 ·
  8. Strata · Q3_K_M Decode 231.73 · Prefill 785.76 ·

Core figures

Decode (avg)
218.51 tok/s
Prefill (avg)
985.54 tok/s
TTFT (avg)
0.97 s
VRAM (avg)
11.76 GB
MTP acceptance rate
0.54
TTFB
1.75 GB
Power draw
288.5 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3.5-4B
Quantization
Q8_0
Framework
Strata
Version
v0.37
Context length
2048 tokens
Batch size
512
GPU layers
5
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)
11.76 GB
OS
Windows 11 23H2
Driver
572.83
CUDA
12.4
Power draw
288.5 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

    v0.37 · Windows 11 23H2 · CUDA 12.4 · 2048 ctx

    218.51 tok/s

    Decode

    985.54 tok/s

    Prefill

    0.97 s

    TTFT

    11.76 GB

    VRAM

    0.54

    MTP

    288.5 W

    Power draw

    L2 batch seed 2026-10-19: Strata on hw_id=14 running model_id=48 at Q8_0