NVIDIA has introduced a new graphics processor in the field of artificial intelligence. This was reported by Zamin.uz.
This new chip is called GB200 and is designed for long-context inference. It has the capability to continuously process large volumes of data and is specifically created for large language models and generative artificial intelligence.
The GB200 chip supports context of up to one million tokens. This allows AI systems to work with multi-page texts and provide logical and accurate responses.
This capability is very important for coding, document analysis, and processing other complex texts. NVIDIA developed this chip based on the GH200 Grace Hopper Superchip platform.
This architecture helps improve the efficiency of large language models, ensures fast inference, and reduces energy consumption. Servers based on the GB200 chip are planned to be released to the market in 2026.
The new technology serves to open new opportunities and enhance efficiency in the field of artificial intelligence.
