Arao AI
Back to search results

elephant-1 · IQ3_S

Start a discussion
NVIDIA H100 80GB HBM380GBStrata v0.37L2 Cross-framework verified

This page aggregates 1 real-world runs of elephant-1 (IQ3_S) on NVIDIA H100 80GB HBM3 with Strata, contributed by 1 independent source platforms; metrics are averages of published measurements.

322.41 tok/s

Decode

Decode speed

1,596.98 tok/s

Prefill

Prefill speed

0.4 s

TTFT

Time to first token

14.57 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

13 days ago

Last verified

Performance

  1. Ollama · Q3_K_M Decode 303.51 · Prefill 1532.77 ·
  2. Strata · IQ3_S (current)Decode 322.41 · Prefill 1596.98 ·

Core figures

Decode (avg)
322.41 tok/s
Prefill (avg)
1,596.98 tok/s
TTFT (avg)
0.4 s
VRAM (avg)
14.57 GB
MTP acceptance rate
0.77
TTFB
2.58 GB
Power draw
587.5 W

Configuration

Member-level fields are taken from the most recent run

Model
elephant-1
Quantization
IQ3_S
Framework
Strata
Version
v0.37
Context length
16384 tokens
Batch size
256
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)
14.57 GB
OS
Ubuntu 22.04
Driver
572.83
CUDA
13.0
Power draw
587.5 W

Get started

Original model resources

Base model resources may not match this quantization. Use the verified deployment weights above when available.

Sources & evidence

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

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-22

    v0.37 · Ubuntu 22.04 · CUDA 13.0 · 16384 ctx

    322.41 tok/s

    Decode

    1,596.98 tok/s

    Prefill

    0.4 s

    TTFT

    14.57 GB

    VRAM

    0.77

    MTP

    587.5 W

    Power draw

    L2 batch seed 2026-10-22: Strata on hw_id=55 running model_id=95 at IQ3_S