Arao AI
Back to search results

DeepSeek-R1-Distill-Qwen-14B · IQ1_S

Start a discussion
AMD Radeon RX 9070 XT16GBOllama v0.47L2 Cross-framework verified

This page aggregates 1 real-world runs of DeepSeek-R1-Distill-Qwen-14B (IQ1_S) on AMD Radeon RX 9070 XT with Ollama, contributed by 1 independent source platforms; metrics are averages of published measurements.

91.31 tok/s

Decode

Decode speed

484.29 tok/s

Prefill

Prefill speed

0.75 s

TTFT

Time to first token

14.45 GB

VRAM

VRAM usage

L2 Cross-framework verified

1

Measured runs

1

Independent sources

Other

Source platforms

5 days ago

Last verified

Performance

  1. llama.cpp · IQ2_XS Decode 150.94 · Prefill 517.81 ·
  2. Ollama · IQ1_S (current)Decode 91.31 · Prefill 484.29 ·

Core figures

Decode (avg)
91.31 tok/s
Prefill (avg)
484.29 tok/s
TTFT (avg)
0.75 s
VRAM (avg)
14.45 GB
MTP acceptance rate
—
TTFB
0.73 GB
Power draw
311.5 W

Configuration

Member-level fields are taken from the most recent run

Model
DeepSeek-R1-Distill-Qwen-14B
Quantization
IQ1_S
Framework
Ollama
Version
v0.47
Context length
16384 tokens
Batch size
256
GPU layers
28
Flash Attention
Off

Hardware

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

GPU
AMD Radeon RX 9070 XT
Nominal VRAM
16 GB
Measured VRAM (avg)
14.45 GB
OS
Ubuntu 24.04
Driver
amdgpu 6.11
CUDA
N/A
Power draw
311.5 W

Sources & evidence

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

  1. L2 Cross-framework verifiedOtherOriginal link Verified on 2026-10-13

    v0.47 · Ubuntu 24.04 · CUDA N/A · 16384 ctx

    91.31 tok/s

    Decode

    484.29 tok/s

    Prefill

    0.75 s

    TTFT

    14.45 GB

    VRAM

    —

    MTP

    311.5 W

    Power draw

    L2 batch seed 2026-10-13: Ollama on hw_id=16 running model_id=71 at IQ1_S