Arao AI
Back to search results

Llama-3.1-8B-Instruct · IQ3_S

Start a discussion
NVIDIA RTX 508016GBStrata v0.26L2 Cross-framework verified

This page aggregates 1 real-world runs of Llama-3.1-8B-Instruct (IQ3_S) on NVIDIA RTX 5080 with Strata, contributed by 1 independent source platforms; metrics are averages of published measurements.

302.96 tok/s

Decode

Decode speed

1,550.76 tok/s

Prefill

Prefill speed

2.09 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. Ollama · Q5_K_S Decode 149.47 · Prefill 563.58 ·
  2. Strata · IQ3_S (current)Decode 302.96 · Prefill 1550.76 ·
  3. vLLM · IQ4_XS Decode 224.27 · Prefill 989.89 ·
  4. Strata · IQ3_XS Decode 308.04 · Prefill 1362.37 ·
  5. Strata · Q3_K_S Decode 323.81 · Prefill 1737.95 ·

Core figures

Decode (avg)
302.96 tok/s
Prefill (avg)
1,550.76 tok/s
TTFT (avg)
2.09 s
VRAM (avg)
13.19 GB
MTP acceptance rate
0.56
TTFB
2.42 GB
Power draw
327 W

Configuration

Member-level fields are taken from the most recent run

Model
Llama-3.1-8B-Instruct
Quantization
IQ3_S
Framework
Strata
Version
v0.26
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 RTX 5080
Nominal VRAM
16 GB
Measured VRAM (avg)
13.19 GB
OS
Windows 11 23H2
Driver
560.94
CUDA
12.4
Power draw
327 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

    v0.26 · Windows 11 23H2 · CUDA 12.4 · 16384 ctx

    302.96 tok/s

    Decode

    1,550.76 tok/s

    Prefill

    2.09 s

    TTFT

    13.19 GB

    VRAM

    0.56

    MTP

    327 W

    Power draw

    L2 batch seed 2026-10-19: Strata on hw_id=14 running model_id=68 at IQ3_S