OpenAI, the research laboratory dedicated to creating artificial general intelligence, has announced new changes to its two primary language models, GPT-3 and GPT-4. The company released updated versions of these models, offering a range of new features and improved performance. The new function calling feature allows software developers to create chatbots that answer questions with external tools, improving plugin and other services that use ChatGPT or the official APIs. Another improvement to GPT-3.5 is that the model now includes more reliable steerability, which gives users more control over its personality in responses, resulting in a more significant impact. A 32k version of GPT-4 is also available through the API, providing extended context length for better comprehension of texts.
Moreover, OpenAI has lowered API access prices, resulting in a 25% reduction in the cost of GPT-3.5 input tokens, at $0.0015 per 1K input tokens and $0.002 per 1K output tokens. In contrast, the more powerful 16K version of GPT-3.5 is priced at $0.003 per 1K input tokens and $0.004 per 1K output tokens, while costs for GPT-4 remain unchanged. Lastly, the waitlist for the GPT-4 API is now open to more people, with hopes of making it available to free ChatGPT users soon.
These new changes are designed to make GPT-3 and GPT-4 more useful for developers and individuals who use OpenAI technology. OpenAI aims to provide better functionality while reducing prices and extending access to its technology, hoping to democratize the benefits of artificial general intelligence. The company continues to work on updates for ChatGPT, with new plugins now in public testing. The latest changes indicate that OpenAI is committed to innovation and growth, making its technology more accessible to more people.
