Arao AI
Back to search results

Llama-3.1-8B-Instruct · IQ3_S

Start a discussion
NVIDIA H2097GBOllama v0.39L2 Cross-framework verified

This page aggregates 1 real-world runs of Llama-3.1-8B-Instruct (IQ3_S) on NVIDIA H20 with Ollama, contributed by 1 independent source platforms; metrics are averages of published measurements.

56.57 tok/s

Decode

Decode speed

208.48 tok/s

Prefill

Prefill speed

2.36 s

TTFT

Time to first token

13.19 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

16 days ago

Last verified

Performance

  1. llama.cpp · Q3_K_M Decode 65.35 · Prefill 226.54 ·
  2. Ollama · IQ3_S (current)Decode 56.57 · Prefill 208.48 ·
  3. Ollama · Q5_K_M Decode 43.67 · Prefill 144.84 ·

Core figures

Decode (avg)
56.57 tok/s
Prefill (avg)
208.48 tok/s
TTFT (avg)
2.36 s
VRAM (avg)
13.19 GB
MTP acceptance rate
—
TTFB
0.45 GB
Power draw
370.5 W

Configuration

Member-level fields are taken from the most recent run

Model
Llama-3.1-8B-Instruct
Quantization
IQ3_S
Framework
Ollama
Version
v0.39
Context length
4096 tokens
Batch size
256
GPU layers
33
Flash Attention
On

Hardware

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

GPU
NVIDIA H20
Nominal VRAM
97 GB
Measured VRAM (avg)
13.19 GB
OS
Windows 11 24H2
Driver
560.94
CUDA
12.4
Power draw
370.5 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

    v0.39 · Windows 11 24H2 · CUDA 12.4 · 4096 ctx

    56.57 tok/s

    Decode

    208.48 tok/s

    Prefill

    2.36 s

    TTFT

    13.19 GB

    VRAM

    —

    MTP

    370.5 W

    Power draw

    L2 batch seed 2026-10-19: Ollama on hw_id=56 running model_id=68 at IQ3_S