Back to search results
Llama-3-8B-Instruct · Q3_K_L
NVIDIA RTX 508016GBStrata v0.32L2 Cross-framework verified
This page aggregates 1 real-world runs of Llama-3-8B-Instruct (Q3_K_L) on NVIDIA RTX 5080 with Strata, contributed by 1 independent source platforms; metrics are averages of published measurements.
297.89 tok/s
Decode
Decode speed
1,445.32 tok/s
Prefill
Prefill speed
2.26 s
TTFT
Time to first token
14.2 GB
VRAM
VRAM usage
L2 Cross-framework verified
1
Measured runs
1
Independent sources
Other
Source platforms
17 days ago
Last verified
Performance
- Ollama · IQ3_S Decode 199.24 · Prefill 946.23 ·
- Ollama · Q3_K_S Decode 204.49 · Prefill 898.01 ·
- Strata · IQ1_S Decode 349.68 · Prefill 1703.4 ·
- Strata · Q3_K_L (current)Decode 297.89 · Prefill 1445.32 ·
- vLLM · Q5_K_M Decode 163.66 · Prefill 702.95 ·
Core figures
- Decode (avg)
- 297.89 tok/s
- Prefill (avg)
- 1,445.32 tok/s
- TTFT (avg)
- 2.26 s
- VRAM (avg)
- 14.2 GB
- MTP acceptance rate
- 0.56
- TTFB
- 2.38 GB
- Power draw
- 259 W
Configuration
Member-level fields are taken from the most recent run
- Model
- Llama-3-8B-Instruct
- Quantization
- Q3_K_L
- Framework
- Strata
- Version
- v0.32
- Context length
- 2048 tokens
- Batch size
- 1024
- GPU layers
- 8
- Flash Attention
- Off
Hardware
Nominal and measured figures are shown side by side; whether it runs is the reader's call
- GPU
- NVIDIA RTX 5080
- Nominal VRAM
- 16 GB
- Measured VRAM (avg)
- 14.2 GB
- OS
- Ubuntu 22.04
- Driver
- 572.83
- CUDA
- 12.6
- Power draw
- 259 W
Sources & evidence
1 measured records in total, each traceable to its original source
v0.32 · Ubuntu 22.04 · CUDA 12.6 · 2048 ctx
297.89 tok/s
Decode
1,445.32 tok/s
Prefill
2.26 s
TTFT
14.2 GB
VRAM
0.56
MTP
259 W
Power draw
L2 batch seed 2026-10-19: Strata on hw_id=14 running model_id=69 at Q3_K_L