Arao AI
Back to search results

DeepSeek-R1-Distill-Qwen-14B · Q4_K_M

Start a discussion
NVIDIA RTX 409024GBllama.cpp b4200L0 Self-reported

This page aggregates 1 real-world runs of DeepSeek-R1-Distill-Qwen-14B (Q4_K_M) on NVIDIA RTX 4090 with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.

92.5 tok/s

Decode

Decode speed

693.8 tok/s

Prefill

Prefill speed

0.14 s

TTFT

Time to first token

9 GB

VRAM

VRAM usage

L0 Self-reported

1

Measured runs

1

Independent sources

Other

Source platforms

Today

Last verified

Performance

  1. vLLM · Q5_K_M Decode 82.4 · Prefill 618 ·
  2. vLLM · Q4_K_M Decode 100.8 · Prefill 756 ·
  3. llama.cpp · Q5_K_M Decode 75.6 · Prefill 567 ·
  4. llama.cpp · Q4_K_M (current)Decode 92.5 · Prefill 693.8 ·

Core figures

Decode (avg)
92.5 tok/s
Prefill (avg)
693.8 tok/s
TTFT (avg)
0.14 s
VRAM (avg)
9 GB
MTP acceptance rate
—
TTFB
— GB
Power draw
— W

Configuration

Member-level fields are taken from the most recent run

Model
DeepSeek-R1-Distill-Qwen-14B
Quantization
Q4_K_M
Framework
llama.cpp
Version
b4200
Context length
4096 tokens
Batch size
512
GPU layers
99
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)
9 GB
OS
Linux
Driver
570.65
CUDA
12.6

Sources & evidence

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

  1. L0 Self-reportedOtherOriginal link Verified on 2026-10-06

    b4200 · Linux · CUDA 12.6 · 4096 ctx

    92.5 tok/s

    Decode

    693.8 tok/s

    Prefill

    0.14 s

    TTFT

    9 GB

    VRAM

    —

    MTP

    — W

    Power draw

    L0 理论估算（add-500-records-batch2）