Arao AI
Back to search results

Llama-3.1-8B-Instruct · Q5_K_M

Start a discussion
NVIDIA H100 80GB HBM380GBStrata v0.53L2 Cross-framework verified

This page aggregates 1 real-world runs of Llama-3.1-8B-Instruct (Q5_K_M) on NVIDIA H100 80GB HBM3 with Strata, contributed by 1 independent source platforms; metrics are averages of published measurements.

362.58 tok/s

Decode

Decode speed

1,954.02 tok/s

Prefill

Prefill speed

0.71 s

TTFT

Time to first token

17.72 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 · IQ1_S Decode 337.53 · Prefill 1795.74 ·
  2. llama.cpp · Q4_K_S Decode 314.55 · Prefill 1364.71 ·
  3. Strata · Q5_K_M (current)Decode 362.58 · Prefill 1954.02 ·

Core figures

Decode (avg)
362.58 tok/s
Prefill (avg)
1,954.02 tok/s
TTFT (avg)
0.71 s
VRAM (avg)
17.72 GB
MTP acceptance rate
0.57
TTFB
2.9 GB
Power draw
667 W

Configuration

Member-level fields are taken from the most recent run

Model
Llama-3.1-8B-Instruct
Quantization
Q5_K_M
Framework
Strata
Version
v0.53
Context length
16384 tokens
Batch size
512
GPU layers
8
Flash Attention
Off

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)
17.72 GB
OS
Ubuntu 24.04
Driver
560.94
CUDA
13.0
Power draw
667 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

    v0.53 · Ubuntu 24.04 · CUDA 13.0 · 16384 ctx

    362.58 tok/s

    Decode

    1,954.02 tok/s

    Prefill

    0.71 s

    TTFT

    17.72 GB

    VRAM

    0.57

    MTP

    667 W

    Power draw

    L2 batch seed 2026-10-19: Strata on hw_id=55 running model_id=68 at Q5_K_M