ChatGPT just added new features and lowered API prices
OpenAI continued to work on updates to ChatGPT, with a new GPT-4 language model and plugins now in public testing. This Tuesday, OpenAI has announced a major update to its API offerings for large language models : GPT-3.5 Turbo (the default model of ChatGPT) and GPT-4 (available in Bing Chat and ChatGPT Pro).
In large language models (LLM), the “pop-up window” is a kind of short-term memory that stores the content of the prompt or, in the case of a chatbot, the entire content of the conversation in course. In language models, increasing context size has become a technology race, with Anthropic recently announcing a 75,000-word popup option for its Claude language model. Also, OpenAI has developed a 32,000 keyword version of GPT-4, but it is not yet publicly available.
In a similar vein, OpenAI just introduced a new version of gpt-3.5-turbo with 16,000 popups, called, unsurprisingly, “gpt-3.5-turbo-16k”, which allows a prompt to have a maximum length of 16,000 tokens. With four times the context length of the standard version (4000), gpt-3.5-turbo-16k can handle around 20 pages of text in a single request. This is a huge step forward for developers who need a model that can process and generate responses for larger chunks of text.
Also, updated models include a new feature called Function callingwhich makes it easier for software developers to create chatbots who answer questions using external tools. This should improve plugins and other services that use ChatGPT or official APIs.
The GPT-3.5 update also includes a “more reliable orientability”, which means she should give you more control over her personality in the responses. Orientability is what makes prompts like “pretend you’re a hacker” work well, and previously worked much better in GPT-4 than in GPT-3.5.
A better price
OpenAI also reduces API access prices. The cost of GPT-3.5 input tokens is now 25% lower, at $0.0015 per 1,000 input tokens and $0.002 per 1,000 output tokens. The more powerful 16K version of GPT-3.5 costs $0.003 per 1K input tokens and $0.004 per 1K output tokens, and GPT-4’s costs don’t seem to have changed.
Finally, OpenAI opens the waiting list for the GPT-4 API to more people. The company hopes to remove the waitlist entirely soon, which could be a sign that GPT-4 will be available to free ChatGPT users in the not-too-distant future.
