Back to search results
DeepSeek-R1-Distill-Qwen-14B · IQ1_S
NVIDIA RTX 509032GBllama.cpp v0.34L2 Cross-framework verified
This page aggregates 1 real-world runs of DeepSeek-R1-Distill-Qwen-14B (IQ1_S) on NVIDIA RTX 5090 with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.
313.16 tok/s
Decode
Decode speed
1,103.72 tok/s
Prefill
Prefill speed
1.14 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
Today
Last verified
Performance
- llama.cpp · IQ1_S (current)Decode 313.16 · Prefill 1103.72 ·
- vLLM · IQ3_XS Decode 290.25 · Prefill 1286.52 ·
- vLLM · Q5_K_S Decode 193.3 · Prefill 1031.83 ·
- vLLM · Q5_K_M Decode 146.5 · Prefill 1098.8 ·
- vLLM · Q4_K_M Decode 179.2 · Prefill 1344 ·
- llama.cpp · Q5_K_M Decode 134.4 · Prefill 1008 ·
- llama.cpp · Q4_K_M Decode 164.4 · Prefill 1233 ·
Core figures
- Decode (avg)
- 313.16 tok/s
- Prefill (avg)
- 1,103.72 tok/s
- TTFT (avg)
- 1.14 s
- VRAM (avg)
- 14.45 GB
- MTP acceptance rate
- —
- TTFB
- 2.51 GB
- Power draw
- 465.4 W
Configuration
Member-level fields are taken from the most recent run
- Model
- DeepSeek-R1-Distill-Qwen-14B
- Quantization
- IQ1_S
- Framework
- llama.cpp
- Version
- v0.34
- Context length
- 32768 tokens
- Batch size
- 256
- GPU layers
- 33
- Flash Attention
- On
Hardware
Nominal and measured figures are shown side by side; whether it runs is the reader's call
- GPU
- NVIDIA RTX 5090
- Nominal VRAM
- 32 GB
- Measured VRAM (avg)
- 14.45 GB
- OS
- Windows 11 23H2
- Driver
- 572.83
- CUDA
- 12.8
- Power draw
- 465.4 W
Sources & evidence
1 measured records in total, each traceable to its original source
v0.34 · Windows 11 23H2 · CUDA 12.8 · 32768 ctx
313.16 tok/s
Decode
1,103.72 tok/s
Prefill
1.14 s
TTFT
14.45 GB
VRAM
—
MTP
465.4 W
Power draw
L2 batch seed 2026-10-18: llama.cpp on hw_id=13 running model_id=71 at IQ1_S