OpenAI,the company behind ChatGPT,introduced a smaller model called GPT-4o Mini on Thursday.This model is touted as being smarter and more cost-effective than the earlier GPT-3.5 Turbo,which was built for simpler tasks like dialogue.

　　OpenAI aims for developers to use GPT-4o Mini to“significantly expand the range of applications built with AI,”according to a blog post.

　　Chatbots like ChatGPT serve as the interface for large language models(LLMs)such as GPT-4o Mini and the original,larger GPT-4o.These models are trained to understand human language and generate human-like content.

　　An LLM can have over a billion parameters,which measure the content it can process before generating a response.While LLMs can learn and understand a lot,they are not always ideal due to their high cost and energy consumption,requiring expansive server farms and cloud access.

　　A smaller language model like GPT-4o Mini offers a compromise by providing AI capabilities and speed without the same computing resource demands or costs.Examples of similar small models include Microsoft’s Phi-3 Mini and Google’s Gemini 1.5 Flash,designed for specific high-volume tasks.

　　The Features of GPT-4o Mini

　　Starting Thursday,both free and paid ChatGPT users can access GPT-4o Mini,replacing GPT-3.5,which was released in November 2022.

　　GPT-4o Mini currently supports text and vision through the OpenAI API,which developers use to build applications with OpenAI technology.Future updates will include support for text,image,video,and audio inputs and outputs.

　　Enterprise users will have access to GPT-4o Mini starting the week of July 22.

　　OpenAI states that GPT-4o Mini excels in mathematical reasoning and coding and has demonstrated skills in reasoning tasks.Financial tech startup Ramp and email app Superhuman have tested GPT-4o Mini for extracting data from files and generating email responses.

　　The new model has a context window of 128,000 tokens,allowing it to remember more within a conversation compared to GPT-3.5 Turbo’s 16,000 tokens.

　　Cost and Efficiency

　　GPT-4o Mini is priced at 15 cents per million input tokens and 60 cents per million output tokens,roughly equivalent to 2,500 book pages.In contrast,GPT-4o,released in May,costs$5 per million input tokens and$2.50 per million output tokens.

　　OpenAI envisions a future where AI models are seamlessly integrated into every app and website,and GPT-4o Mini is paving the way for developers to build and scale AI applications more efficiently and affordably.

　　Safety Measures

　　GPT-4o Mini uses the same safety parameters as GPT-4o.Additionally,it employs a safety technique called instruction hierarchy,prioritizing prompts from developers over third parties to decrease vulnerabilities to external threats,according to The Verge.

　　For more insights on AI,visit our AI Atlas hub,which includes product reviews,news,tips,and explainers.