Arao AI
Back to search results

Qwen3-1.7B · IQ2_M

Start a discussion
NVIDIA RTX 408016GBvLLM v0.31L2 Cross-framework verified

This page aggregates 1 real-world runs of Qwen3-1.7B (IQ2_M) on NVIDIA RTX 4080 with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.

224.82 tok/s

Decode

Decode speed

990.41 tok/s

Prefill

Prefill speed

0.34 s

TTFT

Time to first token

3.96 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

5 days ago

Last verified

Performance

  1. vLLM · IQ2_M (current)Decode 224.82 · Prefill 990.41 ·
  2. vLLM · Q3_K_M Decode 197.14 · Prefill 742.62 ·
  3. Ollama · Q4_K_S Decode 163.88 · Prefill 605 ·
  4. Ollama · Q8_0 Decode 101.94 · Prefill 440.86 ·

Core figures

Decode (avg)
224.82 tok/s
Prefill (avg)
990.41 tok/s
TTFT (avg)
0.34 s
VRAM (avg)
3.96 GB
MTP acceptance rate
—
TTFB
1.8 GB
Power draw
373.5 W

Configuration

Member-level fields are taken from the most recent run

Model
Qwen3-1.7B
Quantization
IQ2_M
Framework
vLLM
Version
v0.31
Context length
16384 tokens
Batch size
16
GPU layers
999
Flash Attention
Off

Hardware

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

GPU
NVIDIA RTX 4080
Nominal VRAM
16 GB
Measured VRAM (avg)
3.96 GB
OS
Ubuntu 24.04
Driver
555.85
CUDA
12.8
Power draw
373.5 W

Sources & evidence

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

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-13

    v0.31 · Ubuntu 24.04 · CUDA 12.8 · 16384 ctx

    224.82 tok/s

    Decode

    990.41 tok/s

    Prefill

    0.34 s

    TTFT

    3.96 GB

    VRAM

    —

    MTP

    373.5 W

    Power draw

    L2 batch seed 2026-10-13: vLLM on hw_id=41 running model_id=55 at IQ2_M