April 25th.NvidiaIt was announced in an official press releaseCodex Depend on OpenAI THE LATEST FRONT-LINE MODEL GPT-5.5 DRIVES AND OPERATES ON THE BRITISH GB200 NVL72 RACK SYSTEM。
Within Weida, Codex has been used to cover multiple teams such as engineering, products, law, marketing, finance, sales, human resources, operations and developers。Over 10000 British Wedda employees are already using GPT-5.5 edition CodexAnd their feedback was that the effect was “shocking” and even “changing life”。
Engineer Young Wai DaA few weeks have passed through Codex using GPT-5.5Efficiency gains are already quantifiable. GPT-5.5 runs on GB200 NVL72. GB200 NVL72 can reduce the cost per million token 35 times and increase the token output by 50 times per megawatt per second compared to the previous generation. It is this cost and energy efficiency advantage that gives frontier model reasoning the opportunity to truly land on the enterprise scale。
The commissioning, which took days to complete, will now be completed in a few hours. Experiments that used to take weeks to advance in complex, multi-document coding can now be seen overnight. Teams can also deliver end-to-end functions directly through natural language tips, which are more reliable than early models and less mobile and back-to-work。
FOUNDER AND CEO OF INWEIDA Jen-Hsun HuangIn a full-time e-mailUrging employees to use CodexI DON'T KNOW. "LET'S JUMP TO LIGHT SPEED AND WELCOME TO THE AI AGE."