OpenAI has announced the release of its latest text-generating AI models, GPT-3.5-turbo and GPT-4, which come equipped with function calling, a feature that aids in software development. Function calling essentially allows developers to describe programming functions to GPT-3.5-turbo and GPT-44, which then generates code to execute these functions. The new models also have a larger context window, measuring 16,000 tokens, compared to the 4,096 tokens of the vanilla version, which improves their ability to generate additional text, reduce forgetting of recent conversations, and avoid problematic (off-topic) responses.
GPT-3.5-turbo has twice the price of the vanilla version but also has four times the context length and can process 20 pages of text in one go, which is more than even Anthropic – its flagship model. OpenAI is currently testing GPT-4 with a 32,000-token context window in limited release. OpenAI is also discounting its popular text embedding model, text-embedding-ada-002, which is used for search and recommendation and uses tokens, hence lowering its cost from $0.0004 per 1,000 tokens to $0.0001 – a 75% reduction.
CEO Sam Altman has confirmed that OpenAI’s focus remains on incremental updates to existing models, despite the successful launch of GPT-4 in March, and that his team is working on a successor to GPT-4. The move to release new models with fresh features, including function calling, is to keep up with generative AI competition and enhance OpenAI’s pricing model.
Frequently Asked Questions (FAQs) Related to the Above News
What are the latest AI models released by OpenAI?
OpenAI has released GPT-3.5-turbo and GPT-4.
What is the function calling feature in GPT-3.5-turbo and GPT-4?
Function calling allows developers to describe programming functions to GPT-3.5-turbo and GPT-4, which then generates code to execute these functions.
How does the larger context window in GPT-3.5-turbo and GPT-4 improve their abilities?
The larger context window, measuring 16,000 tokens, improves their ability to generate additional text, reduce forgetting of recent conversations, and avoid problematic (off-topic) responses.
How much more expensive is GPT-3.5-turbo compared to the vanilla version?
GPT-3.5-turbo has twice the price of the vanilla version.
How much text can GPT-3.5-turbo process in one go?
GPT-3.5-turbo can process 20 pages of text in one go, which is more than even Anthropic – its flagship model.
What is the discount offered by OpenAI on its popular text embedding model?
OpenAI is offering a 75% discount on its popular text embedding model, text-embedding-ada-002, lowering its cost from $0.0004 per 1,000 tokens to $0.0001.
Is OpenAI planning to focus on incremental updates or launch a successor to GPT-4?
CEO Sam Altman has confirmed that OpenAI's focus remains on incremental updates to existing models, despite the successful launch of GPT-4 in March, and that his team is working on a successor to GPT-4.
Why did OpenAI release new models with fresh features?
The move to release new models with fresh features, including function calling, is to keep up with generative AI competition and enhance OpenAI's pricing model.
Please note that the FAQs provided on this page are based on the news article published. While we strive to provide accurate and up-to-date information, it is always recommended to consult relevant authorities or professionals before making any decisions or taking action based on the FAQs or the news article.
