Back to search results
Llama-3.2-3B-Instruct · IQ2_XS
NVIDIA RTX 4080 Super16GBllama.cpp v0.32L2 Cross-framework verified
This page aggregates 1 real-world runs of Llama-3.2-3B-Instruct (IQ2_XS) on NVIDIA RTX 4080 Super with llama.cpp, contributed by 1 independent source platforms; metrics are averages of published measurements.
372.98 tok/s
Decode
Decode speed
1,597.99 tok/s
Prefill
Prefill speed
2.35 s
TTFT
Time to first token
5.12 GB
VRAM
VRAM usage
L2 Cross-framework verified
1
Measured runs
1
Independent sources
Other
Source platforms
17 days ago
Last verified
Performance
- llama.cpp · IQ2_XS (current)Decode 372.98 · Prefill 1597.99 ·
- llama.cpp · IQ4_XS Decode 289.96 · Prefill 1071.96 ·
- Strata · FP16 Decode 277.08 · Prefill 1141.1 ·
- vLLM · IQ3_S Decode 334.5 · Prefill 1159.32 ·
- Strata · Q3_K_L Decode 358.87 · Prefill 1331.79 ·
Core figures
- Decode (avg)
- 372.98 tok/s
- Prefill (avg)
- 1,597.99 tok/s
- TTFT (avg)
- 2.35 s
- VRAM (avg)
- 5.12 GB
- MTP acceptance rate
- —
- TTFB
- 2.98 GB
- Power draw
- 282.5 W
Configuration
Member-level fields are taken from the most recent run
- Model
- Llama-3.2-3B-Instruct
- Quantization
- IQ2_XS
- Framework
- llama.cpp
- Version
- v0.32
- Context length
- 2048 tokens
- Batch size
- 512
- GPU layers
- 20
- Flash Attention
- On
Hardware
Nominal and measured figures are shown side by side; whether it runs is the reader's call
- GPU
- NVIDIA RTX 4080 Super
- Nominal VRAM
- 16 GB
- Measured VRAM (avg)
- 5.12 GB
- OS
- Ubuntu 24.04
- Driver
- 560.94
- CUDA
- 12.4
- Power draw
- 282.5 W
Sources & evidence
1 measured records in total, each traceable to its original source
v0.32 · Ubuntu 24.04 · CUDA 12.4 · 2048 ctx
372.98 tok/s
Decode
1,597.99 tok/s
Prefill
2.35 s
TTFT
5.12 GB
VRAM
—
MTP
282.5 W
Power draw
L2 batch seed 2026-10-19: llama.cpp on hw_id=20 running model_id=70 at IQ2_XS