Arao AI
Back to search results

Qwen3.5-4B · Q2_K

Start a discussion
NVIDIA H100 80GB HBM380GBStrata v0.24L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3.5-4B (Q2_K) on NVIDIA H100 80GB HBM3 with Strata, contributed by 1 independent source platforms; metrics are averages of published measurements.

373.13 tok/s

Decode

Decode speed

1,907.75 tok/s

Prefill

Prefill speed

2.18 s

TTFT

Time to first token

6.72 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 · IQ3_S Decode 329.61 · Prefill 1536.04 ·
  2. vLLM · Q3_K_S Decode 362.27 · Prefill 1696.48 ·
  3. Strata · Q2_K (current)Decode 373.13 · Prefill 1907.75 ·

Core figures

Decode (avg)
373.13 tok/s
Prefill (avg)
1,907.75 tok/s
TTFT (avg)
2.18 s
VRAM (avg)
6.72 GB
MTP acceptance rate
0.62
TTFB
2.99 GB
Power draw
565.5 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3.5-4B
Quantization
Q2_K
Framework
Strata
Version
v0.24
Context length
2048 tokens
Batch size
256
GPU layers
15
Flash Attention
On

Hardware

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

GPU
NVIDIA H100 80GB HBM3
Nominal VRAM
80 GB
Measured VRAM (avg)
6.72 GB
OS
Ubuntu 24.04
Driver
560.94
CUDA
12.8
Power draw
565.5 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

    v0.24 · Ubuntu 24.04 · CUDA 12.8 · 2048 ctx

    373.13 tok/s

    Decode

    1,907.75 tok/s

    Prefill

    2.18 s

    TTFT

    6.72 GB

    VRAM

    0.62

    MTP

    565.5 W

    Power draw

    L2 batch seed 2026-10-19: Strata on hw_id=55 running model_id=48 at Q2_K