Arao AI
Back to search results

Llama-3.2-3B-Instruct · IQ2_M

Start a discussion
NVIDIA RTX 4070 Ti12GBStrata v0.36L2 Cross-framework verified

This page aggregates 1 real-world runs of Llama-3.2-3B-Instruct (IQ2_M) on NVIDIA RTX 4070 Ti with Strata, contributed by 1 independent source platforms; metrics are averages of published measurements.

347.68 tok/s

Decode

Decode speed

1,284.54 tok/s

Prefill

Prefill speed

0.55 s

TTFT

Time to first token

5.38 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

Today

Last verified

Performance

  1. llama.cpp · FP16 Decode 157.49 · Prefill 748.12 ·
  2. Strata · IQ2_M (current)Decode 347.68 · Prefill 1284.54 ·
  3. Strata · Q4_K_M Decode 284.07 · Prefill 1145.96 ·

Core figures

Decode (avg)
347.68 tok/s
Prefill (avg)
1,284.54 tok/s
TTFT (avg)
0.55 s
VRAM (avg)
5.38 GB
MTP acceptance rate
0.63
TTFB
2.78 GB
Power draw
275.1 W

Configuration

Member-level fields are taken from the most recent run

Model
Llama-3.2-3B-Instruct
Quantization
IQ2_M
Framework
Strata
Version
v0.36
Context length
32768 tokens
Batch size
1024
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 4070 Ti
Nominal VRAM
12 GB
Measured VRAM (avg)
5.38 GB
OS
Ubuntu 24.04
Driver
565.57
CUDA
12.4
Power draw
275.1 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 12.4 · 32768 ctx

    347.68 tok/s

    Decode

    1,284.54 tok/s

    Prefill

    0.55 s

    TTFT

    5.38 GB

    VRAM

    0.63

    MTP

    275.1 W

    Power draw

    L2 batch seed 2026-10-18: Strata on hw_id=42 running model_id=70 at IQ2_M