Arao AI
Back to search results

Llama-3-8B-Instruct · IQ1_S

Start a discussion
NVIDIA RTX 3060 12GB12GBStrata v0.37L2 Cross-framework verified

This page aggregates 1 real-world runs of Llama-3-8B-Instruct (IQ1_S) on NVIDIA RTX 3060 12GB with Strata, contributed by 1 independent source platforms; metrics are averages of published measurements.

126.53 tok/s

Decode

Decode speed

600.27 tok/s

Prefill

Prefill speed

2.07 s

TTFT

Time to first token

9.16 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

Today

Last verified

Performance

  1. Strata · IQ1_S (current)Decode 126.53 · Prefill 600.27 ·
  2. Strata · Q3_K_S Decode 98.81 · Prefill 492.63 ·

Core figures

Decode (avg)
126.53 tok/s
Prefill (avg)
600.27 tok/s
TTFT (avg)
2.07 s
VRAM (avg)
9.16 GB
MTP acceptance rate
0.7
TTFB
1.01 GB
Power draw
141.9 W

Configuration

Member-level fields are taken from the most recent run

Model
Llama-3-8B-Instruct
Quantization
IQ1_S
Framework
Strata
Version
v0.37
Context length
16384 tokens
Batch size
256
GPU layers
8
Flash Attention
On

Hardware

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

GPU
NVIDIA RTX 3060 12GB
Nominal VRAM
12 GB
Measured VRAM (avg)
9.16 GB
OS
Ubuntu 22.04
Driver
560.94
CUDA
12.4
Power draw
141.9 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

    v0.37 · Ubuntu 22.04 · CUDA 12.4 · 16384 ctx

    126.53 tok/s

    Decode

    600.27 tok/s

    Prefill

    2.07 s

    TTFT

    9.16 GB

    VRAM

    0.7

    MTP

    141.9 W

    Power draw

    L2 batch seed 2026-10-18: Strata on hw_id=15 running model_id=69 at IQ1_S