Live Floor
n/a
No active offers
GPU Profile24 GB VRAMData Center GPU
K80
The NVIDIA K80 features the Kepler architecture with 24 GB of GDDR5 VRAM split across two GPU dies.
Currently out of stock across tracked providers
Average Price
n/a
Waiting for new inventory
Reliability
n/a
No reliability telemetry
Last Day Delta
n/a
No history yet
30-Day Price Trend
Daily Lowest Price
No price history available yet. Data will appear after the first full day of tracking.
No offers currently available for K80. Check back later or browse other GPU models.
Browse All GPUsGPU Overview
About K80
The NVIDIA K80 features the Kepler architecture with 24 GB of GDDR5 VRAM split across two GPU dies. It is a legacy datacenter accelerator that remains available on budget cloud instances, offering large aggregate memory capacity at minimal cost.
With 24 GB of total VRAM, the K80 can load Large Language Models (LLM) for basic inference tasks, though its throughput is significantly limited by the older Kepler architecture. It is best suited for educational use, initial prototyping, and scenarios where compute cost is the primary constraint.
Key Features
- VRAM24 GB
- CategoryData Center GPU
- Benchmark Score10
Best For
LLM InferenceAI TrainingScalable Deployment