OpenAI is enhancing its text generation models and reducing prices in response to the increasing competition in the generative AI field. In their latest announcement, OpenAI introduced upgraded versions of GPT-3.5-turbo and GPT-4. GPT-4 incorporates a feature called function calling, enabling developers to describe programming functions to the models, which then generate the corresponding code to execute those functions. This functionality has various applications, such as creating chatbots that utilize external tools to answer questions, converting natural language into database queries, and extracting structured data from text.
Apart from function calling, OpenAI is also introducing an expanded context window in a variant of GPT-3.5-turbo. The context window refers to the amount of text considered by the model before generating additional text. Models with smaller context windows tend to forget recent conversations, resulting in digressions or problematic responses. The new GPT-3.5-turbo offers four times the context length, allowing it to consider up to 16,000 tokens of text. However, this version comes at twice the price, with $0.003 per 1,000 input tokens and $0.004 per 1,000 output tokens. OpenAI is currently testing GPT-4 with a 32,000-token context window in limited release.
In terms of pricing adjustments, OpenAI is reducing the cost of GPT-3.5-turbo (the original version without the expanded context window) by 25%. Developers can now utilize the model at $0.0015 per 1,000 input tokens and $0.002 per 1,000 output tokens, equivalent to approximately 700 pages per dollar. Furthermore, OpenAI is also lowering the pricing of text-embedding-ada-002, a popular text embedding model used for search and recommendations. The new price is $0.0001 per 1,000 tokens, reflecting a 75% reduction facilitated by improved system efficiency.
OpenAI has indicated that it is focused on incremental updates to existing models rather than developing entirely new models from scratch. Despite releasing GPT-4 earlier in the year, OpenAI’s CEO, Sam Altman, confirmed that the successor to GPT-4 is not yet in training, suggesting that the company has significant work ahead before beginning work on the next model. OpenAI remains committed to research and infrastructure development, with substantial investments in these areas.