Qwen3-8B · Q5_K_M
This page aggregates 2 real-world runs of Qwen3-8B (Q5_K_M) on NVIDIA RTX 5090 with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.
221.44 tok/s
Decode
Decode speed
1,381.14 tok/s
Prefill
Prefill speed
0.7 s
TTFT
Time to first token
12.06 GB
VRAM
VRAM usage
2
Measured runs
1
Independent sources
Other
Source platforms
2026-10-18
Last verified
Performance
- Ollama · IQ3_S Decode 275.39 · Prefill 1404.89 ·
- Strata · IQ4_XS Decode 340.69 · Prefill 1198.68 ·
- llama.cpp · IQ4_XS Decode 256.42 · Prefill 1150.82 ·
- llama.cpp · Q5_K_M (current)Decode 221.44 · Prefill 1381.14 ·
- vLLM · IQ3_S Decode 285.95 · Prefill 1301.24 ·
- llama.cpp · Q6_K Decode 184.6 · Prefill 725.28 ·
- vLLM · IQ2_M Decode 325.14 · Prefill 1583.07 ·
- vLLM · Q2_K Decode 357.71 · Prefill 1294.78 ·
- vLLM · IQ4_XS Decode 252.51 · Prefill 829.1 ·
- Ollama · Q5_K_M Decode 172.9 · Prefill 940.17 ·
- llama.cpp · Q8_0 Decode 162.5 · Prefill 1218.8 ·
- llama.cpp · Q4_K_M Decode 285.3 · Prefill 2139.8 ·
Core figures
- Decode (avg)
- 221.44 tok/s
- Prefill (avg)
- 1,381.14 tok/s
- TTFT (avg)
- 0.7 s
- VRAM (avg)
- 12.06 GB
- MTP acceptance rate
- —
- TTFB
- 1.65 GB
- Power draw
- 453.4 W
Configuration
Member-level fields are taken from the most recent run
- Model
- Qwen3-8B
- Quantization
- Q5_K_M
- Framework
- llama.cpp
- Version
- v0.44
- Context length
- 4096 tokens
- Batch size
- 256
- 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 5090
- Nominal VRAM
- 32 GB
- Measured VRAM (avg)
- 12.06 GB
- OS
- Ubuntu 24.04
- Driver
- 572.83
- CUDA
- 12.4
- Power draw
- 453.4 W
Sources & evidence
2 measured records in total, each traceable to its original source
v0.44 · Ubuntu 24.04 · CUDA 12.4 · 4096 ctx
205.97 tok/s
Decode
985.47 tok/s
Prefill
1.34 s
TTFT
17.72 GB
VRAM
—
MTP
453.4 W
Power draw
L2 batch seed 2026-10-18: llama.cpp on hw_id=13 running model_id=53 at Q5_K_M
b4200 · Linux · CUDA 12.6 · 4096 ctx
236.9 tok/s
Decode
1,776.8 tok/s
Prefill
0.06 s
TTFT
6.4 GB
VRAM
—
MTP
— W
Power draw
L0 理论估算(add-1000-records-batch3)