OpenAI launches major update to GPT chatbot API for developers and lowers price

News on June 14th, OpenAI today announced a major update to its large language model API (including GPT-4 and gpt-3.5-turbo), including new function calling functions, Reduce usage costs and provide a 16,000 token version for the gpt-3.5-turbo model.

Large language model (LLM) is an artificial intelligence technology that can process natural language. Its "context window" is equivalent to a short-term memory that can store input content or chat robot dialogue. content. In language models, increasing the context window size has become a technical race, and Anthropic recently announced that its Claude language model can provide 75,000 token context window options. In addition, OpenAI has also developed a 32,000 token GPT-4 version, but it has not yet been publicly launched.

OpenAI has just launched a new 16000 context window version of gpt-3.5-turbo, named "gpt-3.5-turbo-16k", which can handle inputs up to 16000 tokens in length, which means that Processing approximately 20 pages of text at a time, this is a big improvement for developers who need to model processing and generate larger blocks of text.

In addition to this change, OpenAI also lists at least four other major new features:

• Introducing function calling capabilities in the Chat Completions API
• Improved and "more steerable" versions of #GPT-4 and gpt-3.5-turbo
• Reduced the price of the "ada" embedded model by 75%
• Reduced gpt -The input token price of the 3.5-turbo model has been reduced by 25%

The function call function makes it easier for developers to build functions that can call external tools, convert natural language into external API calls, or perform database queries. chatbot. For example, it can convert an input like "Send an email to Anya to see if she wants coffee next Friday" into a function call like "send_email (to: string, body: string)". In particular, this feature also makes it easier for API users to generate output in JSON format, something that was previously difficult to achieve.

Regarding the "more controllable" aspect, which is a technical term for how to make LLM behave the way you want, OpenAI said its new "gpt-3.5-turbo-0613" model will include " More reliable control via system messages" functionality. System messages are a special instruction input in the API that tell the model how to behave, such as "You are Grimes and you only talk about milkshakes."

In addition to functional improvements, OpenAI also provides considerable cost reduction. Notably, the token price of the popular gpt-3.5-turbo model is reduced by 25%. This means that developers can now use this model for about $0.0015 per 1000 tokens and $0.002 per 1000 tokens, which equates to about 700 pages of text per dollar. The gpt-3.5-turbo-16k model is priced at US$0.003 per 1000 tokens and US$0.004 per 1000 tokens.

In addition, IT House noticed that OpenAI also reduced the price of its "text-embedding-ada-002" embedding model by 75%. Embedding models are a technique that allows computers to understand words and concepts, converting natural language into a digital language that machines can understand, which is important for tasks such as searching for text and recommending relevant content.

Because OpenAI is constantly updating its models, old models won’t persist. Today, the company also announced it is starting to retire some earlier versions of its models, including gpt-3.5-turbo-0301 and gpt-4-0314. Developers can continue to use these models until September 13, after which these older models will no longer be available, the company said. It is worth noting that OpenAI’s GPT-4 API is still on the waiting list and is not yet fully open.

The above is the detailed content of OpenAI launches major update to GPT chatbot API for developers and lowers price. For more information, please follow other related articles on the PHP Chinese website!