Arao AI
Back to search results

Ministral-3-8B-Instruct-2512 · IQ2_M

Start a discussion
NVIDIA RTX 409024GBStrata v0.38L2 Cross-framework verified

This page aggregates 1 real-world runs of Ministral-3-8B-Instruct-2512 (IQ2_M) on NVIDIA RTX 4090 with Strata, contributed by 1 independent source platforms; metrics are averages of published measurements.

230.86 tok/s

Decode

Decode speed

912.94 tok/s

Prefill

Prefill speed

2.17 s

TTFT

Time to first token

10.84 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

11 days ago

Last verified

Performance

  1. Ollama · Q3_K_S Decode 142.08 · Prefill 478.74 ·
  2. Strata · IQ2_M (current)Decode 230.86 · Prefill 912.94 ·

Core figures

Decode (avg)
230.86 tok/s
Prefill (avg)
912.94 tok/s
TTFT (avg)
2.17 s
VRAM (avg)
10.84 GB
MTP acceptance rate
0.74
TTFB
1.85 GB
Power draw
368 W

Configuration

Member-level fields are taken from the most recent run

Model
Ministral-3-8B-Instruct-2512
Quantization
IQ2_M
Framework
Strata
Version
v0.38
Context length
8192 tokens
Batch size
512
GPU layers
15
Flash Attention
On

Hardware

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

GPU
NVIDIA RTX 4090
Nominal VRAM
24 GB
Measured VRAM (avg)
10.84 GB
OS
Windows 11 24H2
Driver
565.57
CUDA
12.8
Power draw
368 W

Sources & evidence

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

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-24

    v0.38 · Windows 11 24H2 · CUDA 12.8 · 8192 ctx

    230.86 tok/s

    Decode

    912.94 tok/s

    Prefill

    2.17 s

    TTFT

    10.84 GB

    VRAM

    0.74

    MTP

    368 W

    Power draw

    L2 batch seed 2026-10-24: Strata on hw_id=1 running model_id=74 at IQ2_M