한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

a tweet posted by sam altman on the the cost per token has dropped by 98%, the number of tokens in our system has increased by 50 times, model intelligence has made excellent progress, and there is still a little drama in the process. i am looking forward to the progress from this devday to the next.”

indeed, this year's dev day did not disappoint. openai's developer day did not release any new models, but focused on api improvements. this developer day brings a series of important tool updates for developers such as real-time api, prompt caching, model distillation, visual fine-tuning, playground optimization, as well as expanding the scope of o1 api and increasing the rate limit. the pricing of these new apis is not cheap. many developers comment that the combination of pricing and capabilities makes them attractive.

however, the most eye-catching feature is the realtime api - it allows developers to call the underlying model gpt-4o-realtime-preview of chatgpt's advanced speech mode to build a fast and natural speech-to-speech conversation experience in their applications. supports 6 preset voices to achieve low-latency voice interaction and complete the entire conversation process at once.

at the scene, the staff demonstrated the voice assistant built with real-time api, helping more than 100 developers to "call to place an order for 400 strawberries for delivery". this is one of the most fascinating highlights of developer day. developers are amazed by its smooth voice interaction experience, as if they are in a virtual world.

this is not just a technology update, but also a reflection on the future and the development direction of ai.

[keyword] 【wedding dress, dress】