Nvidia CEO Jensen Huang has made a big claim about OpenAI’s latest AI model GPT-6 Astra. After Astra’s launch last week , Huang said AGI has arrived and congratulated the OpenAI team. Huang also revealed that GPT-6 Astra was trained using more than 100,000 Nvidia Grace Blackwell NVLink72 systems.
In a post on X, Huang looked at how quickly OpenAI’s AI models have improved over the last few years.
“GPT-6 Astra, trained on ~100K+ NVIDIA Grace Blackwell NVLink72. From ChatGPT to o1 to Astra in 4 years. AGI has arrived. Congratulations @OpenAI team.” Huang wrote.
He also said that another 400,000 GPUs are coming online next showing just how quickly the demand for AI computing is growing.
AGI stands for Artificial General Intelligence In simple words, it means an AI system that can understand and perform a wide range of tasks at a level similar to, or better than, humans.
Huang’s comment is important because he is one of the biggest names in the AI industry and Nvidia’s chips are powering many of today’s advanced AI models.
However, Huang saying that AGI has arrived does not mean the entire AI industry agrees with him.
OpenAI has described GPT-6 Astra as its most advanced model, with stronger abilities in areas such as computer use, coding, science, cybersecurity and professional work.
The model can also handle complex tasks, follow multiple steps and complete more work with less human help. But whether this is actually AGI is still a matter of debate.
Huang’s post also gives a glimpse of what could come next.
The Nvidia CEO said 400,000 more GPUs are coming online, which could give AI companies even more computing power to train and run advanced models.
For OpenAI, Astra marks another major step in its AI race with companies such as Google and Anthropic.
For Nvidia, the development is equally important. More powerful AI models mean more demand for the company’s chips and data centre systems.