Back to search results
Llama-3-8B-Instruct · Q4_K_S
NVIDIA RTX 409024GBvLLM v0.32L2 Cross-framework verified
This page aggregates 1 real-world runs of Llama-3-8B-Instruct (Q4_K_S) on NVIDIA RTX 4090 with vLLM, contributed by 1 independent source platforms; metrics are averages of published measurements.
157.4 tok/s
Decode
Decode speed
646.64 tok/s
Prefill
Prefill speed
1.85 s
TTFT
Time to first token
15.2 GB
VRAM
VRAM usage
L2 Cross-framework verified
1
Measured runs
1
Independent sources
Other
Source platforms
16 days ago
Last verified
Performance
- llama.cpp · Q5_K_S Decode 118.71 · Prefill 394.1 ·
- vLLM · IQ3_S Decode 177.41 · Prefill 791.84 ·
- vLLM · Q4_K_S (current)Decode 157.4 · Prefill 646.64 ·
- llama.cpp · Q4_K_M Decode 146.97 · Prefill 902.58 ·
- llama.cpp · Q4_K_S Decode 134.4 · Prefill 594.59 ·
- Strata · Q2_K Decode 254.93 · Prefill 1284 ·
- llama.cpp · Q8_0 Decode 91.4 · Prefill 685.5 ·
- llama.cpp · Q5_K_M Decode 133.3 · Prefill 999.8 ·
Core figures
- Decode (avg)
- 157.4 tok/s
- Prefill (avg)
- 646.64 tok/s
- TTFT (avg)
- 1.85 s
- VRAM (avg)
- 15.2 GB
- MTP acceptance rate
- —
- TTFB
- 1.26 GB
- Power draw
- 396.8 W
Configuration
Member-level fields are taken from the most recent run
- Model
- Llama-3-8B-Instruct
- Quantization
- Q4_K_S
- Framework
- vLLM
- Version
- v0.32
- Context length
- 2048 tokens
- Batch size
- 16
- GPU layers
- 999
- 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)
- 15.2 GB
- OS
- Windows 11 23H2
- Driver
- 565.57
- CUDA
- 12.6
- Power draw
- 396.8 W
Sources & evidence
1 measured records in total, each traceable to its original source
v0.32 · Windows 11 23H2 · CUDA 12.6 · 2048 ctx
157.4 tok/s
Decode
646.64 tok/s
Prefill
1.85 s
TTFT
15.2 GB
VRAM
—
MTP
396.8 W
Power draw
L2 batch seed 2026-10-19: vLLM on hw_id=1 running model_id=69 at Q4_K_S