OpenAI Releases New Models and Strikes Content Deal with News Group

Recently, OpenAI has made a series of significant updates and announcements regarding the launch of new models, pricing adjustments, and a partnership with a news group.

Firstly, OpenAI has introduced a new generation of embedding models, including text-embedding-3-small and text-embedding-3-large, along with the introduction of new API usage management tools. These new models have demonstrated excellent performance across multiple benchmark tests and offer more flexible usage options, such as shortening embedding vector dimensions through the dimensions parameter to strike a balance between performance and cost.

Simultaneously, OpenAI has announced an updated version of GPT-3.5 Turbo, named gpt-3.5-turbo-0125, which brings significant improvements in response accuracy and support for non-English languages, along with a 50% price reduction. Additionally, a new preview version of GPT-4 Turbo has been launched, increasing the context window size to 128k tokens and addressing some of the “laziness” issues encountered in generation tasks.

It is worth noting that there are rumors circulating about the release of GPT-4.5 Turbo in June, which is expected to deliver significant improvements in performance and speed. While this news has not been officially confirmed, it has garnered widespread attention in the industry.