What is GPT-4o mini
GPT-4o mini is a newly launched AI large model from OpenAI, offering a more efficient and more cost-effective solution. This new model replaces GPT-3.5 and becomes the most cost-effective small model on the market. GPT-4o mini achieved a high score of 82% on the MMLU benchmark and outperformed GPT-4 in LMSYS's chat preference test. The pricing of this model is extremely competitive, at only 15 cents per million input tokens and 60 cents per million output tokens, an order of magnitude lower in price than previous frontier models, and more than 60% cheaper than GPT-3.5 Turbo.
GPT-4o mini not only supports text input but also supports visual input, and plans to expand in the future to image, video, and audio input and output. The model has a 128K token context window and can process up to 16K output tokens per request, with knowledge coverage up to October 2023. In addition, an improved tokenizer makes it more cost-effective when processing non-English text.
In text intelligence and multimodal reasoning, GPT-4o mini performs excellently, supports multiple languages, and performs well on function calling, mathematical reasoning, and coding tasks. Starting July 19, free users, Plus users, and Team users will be able to access GPT-4o mini, replacing GPT-3.5, and enterprise users will also begin experiencing this new model next week.
Features of GPT-4o mini
- Text intelligence:GPT-4o mini performs excellently in text understanding and generation, and can handle complex text tasks such as natural language understanding, text summarization, content creation, and more.
- Multimodal reasoning：Supports text and visual input, and will expand in the future to image, video, and audio input and output; the model can understand and process multiple types of data.
- Mathematical reasoning：Performs excellently in mathematical problem solving and logical reasoning, and can handle complex mathematical problems and logical reasoning tasks.
- Coding tasks：Coding performance tests show excellent performance, and it can understand and generate code to assist software development and programming tasks.
- Cost-effectiveness：The pricing is extremely competitive, at 15 cents per million input tokens and 60 cents per million output tokens, reducing the development and operating costs of AI applications.
- Easy to access：Free users, Plus users, and Team users can all access GPT-4o mini, and users can conveniently use the most advanced AI technology.
How to use GPT-4o mini
GPT-4o mini can be used for free. To use GPT-4o mini, you only need to visit their official website (https://chat.openai.com/chat), no download requiredbut magic is required, go to the ChatGPT page and log in, and you can choose to register with email or log in with your Google or Microsoft account.
ChatGPT's web interface is easy for all users to use. The display interface includes a text box for users to enter queries and an area that displays results. After entering your text prompt, you will then receive the response information returned by ChatGPT.