news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

IT Home reported on July 22 that on July 18,OpenAI Released a new model called "the most cost-effective small model"GPT-4o Mini. CEO Altman then launched the “cargo mode” on X (Twitter): “15 cents per million input tokens, 60 cents per million output tokens, 82% MMLU, and fast speed. Most importantly, we think people will really enjoy using this new model.”

Most netizens praised it highly in the comment section, but some netizens also thought thatGPT It’s time to change the name of the series model. In response to Altman’s tweet, the netizen said, “You really needChange your name。”

Altman responded in the affirmative.Haha, yes we do.”

The GPT-4o mini model is part of OpenAI's long-term plan to make artificial intelligence more "accessible" because it is much cheaper than previous models while still performing ChatGPT Well-known common tasks includeCustomer support, translation and data processing。

IT Home found that in addition to the GPT series of products, OpenAI has chosen a "unique" naming method for other products. For example, the Japanese pronunciation of the Chinese character "空" isVincent video product "Sora", and Vincent products that are said to be a tribute to the character "Wall-E" from "Wall-E" and the Spanish painter Salvador Dali“DALL-E”. In addition, there is the "Q*" technology and the mysterious advanced AI project that was just exposed recently - its internal code name is "Strawberry".