DeepSeek-R1-Distill-Qwen-14B · Q4_K_M
This page aggregates 2 real-world runs of DeepSeek-R1-Distill-Qwen-14B (Q4_K_M) on NVIDIA RTX Spark with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.
32.88 tok/s
Decode
Decode speed
122.17 tok/s
Prefill
Prefill speed
1.59 s
TTFT
Time to first token
26.21 GB
VRAM
VRAM usage
2
Measured runs
1
Independent sources
Other
Source platforms
16 days ago
Last verified
Performance
- llama.cpp · Q4_K_M (current)Decode 32.88 · Prefill 122.17 ·
- Ollama · IQ3_S Decode 36.95 · Prefill 186.83 ·
- Ollama · Q4_K_M Decode 32.3 · Prefill 137.42 ·
Core figures
- Decode (avg)
- 32.88 tok/s
- Prefill (avg)
- 122.17 tok/s
- TTFT (avg)
- 1.59 s
- VRAM (avg)
- 26.21 GB
- MTP acceptance rate
- —
- TTFB
- 0.26 GB
- Power draw
- 66.35 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
- v0.28
- Context length
- 2048 tokens
- Batch size
- 512
- GPU layers
- 20
- Flash Attention
- Off
Hardware
Nominal and measured figures are shown side by side; whether it runs is the reader's call
- GPU
- NVIDIA RTX Spark
- Nominal VRAM
- 128 GB
- Measured VRAM (avg)
- 26.21 GB
- OS
- Windows 11 24H2
- Driver
- NVIDIA 590.10
- CUDA
- 13.0
- Power draw
- 68 W
- Reference price
- $2,399 · updated 2026-10-19
Prices are not live; check the store page
Sources & evidence
2 measured records in total, each traceable to its original source
v0.28 · Windows 11 24H2 · CUDA 13.0 · 2048 ctx
31.68 tok/s
Decode
118.98 tok/s
Prefill
1.86 s
TTFT
26.21 GB
VRAM
—
MTP
68 W
Power draw
L2 batch seed 2026-10-19: llama.cpp on hw_id=110 running model_id=71 at Q4_K_M
v0.43 · Windows 11 24H2 · CUDA 13.0 · 4096 ctx
34.07 tok/s
Decode
125.36 tok/s
Prefill
1.31 s
TTFT
26.21 GB
VRAM
—
MTP
64.7 W
Power draw
L2 batch seed 2026-10-19: llama.cpp on hw_id=110 running model_id=71 at Q4_K_M