Arao AI
Back to search results

Llama-3-8B-Instruct · IQ3_S

Start a discussion
NVIDIA RTX 5070 Ti16GBStrata v0.36L2 Cross-framework verified

This page aggregates 1 real-world runs of Llama-3-8B-Instruct (IQ3_S) on NVIDIA RTX 5070 Ti with Strata, contributed by 1 independent source platforms; metrics are averages of published measurements.

180.15 tok/s

Decode

Decode speed

920.56 tok/s

Prefill

Prefill speed

1.83 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

Today

Last verified

Performance

  1. vLLM · Q2_K Decode 145.89 · Prefill 573.9 ·
  2. Strata · IQ3_S (current)Decode 180.15 · Prefill 920.56 ·

Core figures

Decode (avg)
180.15 tok/s
Prefill (avg)
920.56 tok/s
TTFT (avg)
1.83 s
VRAM (avg)
13.19 GB
MTP acceptance rate
0.54
TTFB
1.44 GB
Power draw
228.4 W

Configuration

Member-level fields are taken from the most recent run

Model
Llama-3-8B-Instruct
Quantization
IQ3_S
Framework
Strata
Version
v0.36
Context length
2048 tokens
Batch size
1024
GPU layers
15
Flash Attention
Off

Hardware

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

GPU
NVIDIA RTX 5070 Ti
Nominal VRAM
16 GB
Measured VRAM (avg)
13.19 GB
OS
Ubuntu 24.04
Driver
560.94
CUDA
12.8
Power draw
228.4 W

Sources & evidence

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

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-18

    v0.36 · Ubuntu 24.04 · CUDA 12.8 · 2048 ctx

    180.15 tok/s

    Decode

    920.56 tok/s

    Prefill

    1.83 s

    TTFT

    13.19 GB

    VRAM

    0.54

    MTP

    228.4 W

    Power draw

    L2 batch seed 2026-10-18: Strata on hw_id=17 running model_id=69 at IQ3_S