NVIDIA Vera Rubin NVL72 delivers 30x more agentic work per watt
NVIDIA's Vera Rubin NVL72 systems deliver up to 30x higher throughput per megawatt than GB300 NVL72 on agentic workloads, measured with SemiAnalysis AgentX. Agentic requests consume 15x more tokens than chat, and token cost is 35x lower.
How this story unfolded
2 days · 3 reports · 3 of 4 shown
- Aug 22
- Aug 24
NVIDIA by email
Get an email when NVIDIA has news
No news that day, no email.
More stories today
- Study: ChatGPT plus critical-thinking training boosts student performance
- MiniMax H3 generates fake speedpaint timelapses
- Deepgram adds enhanced metrics to Amazon SageMaker AI observability
- NVIDIA MPS on EC2 cuts ASR inference costs by 75%
- Claude diagnoses GPU flaw and builds guard