OpenAI has released new models, including an updated GPT-4 Turbo preview model, and announced a price reduction for GPT-3.5 Turbo API access, responding to user demand. The company also introduced new ways for developers to manage API keys and understand API usage.
The updated GPT-4 Turbo is designed to handle tasks like code generation more comprehensively than its predecessor and aims to reduce instances of the model failing to complete tasks. OpenAI also unveiled the new gpt-3.5-turbo-0125 model and announced a significant price cut for GPT-3.5 Turbo, reducing prices by 50% per thousand tokens and output prices by 25%.
Complaints emerged among ChatGPT users in December 2023 about the bot frequently refusing tasks, which was attributed to the lack of updates to GPT-4. GPT-4, using data available before September 2021, may still exhibit “laziness,” while GPT-4 Turbo has been trained on information as recent as April 2023.
OpenAI introduced smaller AI models known as embeddings, which represent concepts in content like language or code as numerical sequences. Embeddings help computers understand and utilize written text more effectively by converting words and sentences into a format that computers can process.
There are two new models utilizing these embeddings: text-embedding-3-small and the more powerful text-embedding-3-large. These models, indicative of their capacity, can now be used in applications that require efficient retrieval and use of information, with the larger model offering more comprehensive translation and understanding of text.