OpenAI inference cost reduction cut ChatGPT guest traffic from tens of thousands of Nvidia GPUs to just a couple hundred, using software optimization alone. Engineers achieved more than 50% savings ...

One ChatGPT query consumes energy equivalent to running a 40W mini cooling fan for about three minutes. Similarly, a single query uses the same amount of energy as charging your phone with a 5W ...

OpenAI inference cost reduction cut ChatGPT guest traffic from tens of thousands of Nvidia GPUs to just a couple hundred, ...