OpenAI has launched a new AI model called GPT-4o mini. This model is designed to make artificial intelligence more accessible and cost-effective. GPT-4o mini offers impressive capabilities at a fraction of the cost of previous models.
The pricing for GPT-4o mini is very competitive. It costs only 15 cents per million input tokens and 60 cents per million output tokens. This makes it much cheaper than earlier models like GPT-3.5 Turbo.
GPT-4o mini has shown excellent performance in various benchmarks. It scored 82% on the MMLU benchmark. It also outperformed GPT-41 on the LMSYS leaderboard for chat preferences.
The model is versatile and can handle many different tasks. Its low cost and quick response times make it ideal for applications that need multiple model calls. It’s also great for handling large amounts of context or real-time text interactions.
Currently, GPT-4o mini can work with text and vision inputs. OpenAI plans to add support for image, video, and audio inputs and outputs in the future. The model has a large context window of 128K tokens. It can produce up to 16K output tokens per request.
GPT-4o mini’s knowledge is up to date until October 2023. Its improved tokenizer makes it more efficient at handling non-English text. This further reduces costs for users working with multiple languages.
The model excels in both academic and practical applications. It outperforms other small models in areas like reasoning, math, and coding. For example, it scored 87% in mathematical reasoning and 87.2% in coding performance on respected benchmarks.
Safety is a top priority for OpenAI. They have built robust safety measures into GPT-4o mini. These include filtering out harmful content during pre-training. The model also uses reinforcement learning with human feedback to align its behavior with safety policies.
OpenAI has thoroughly tested GPT-4o mini. More than 70 external experts have examined the model. They worked to identify and reduce potential risks, ensuring the model is reliable and safe to use.
GPT-4o mini is now available through several OpenAI APIs. These include the Assistants API, Chat Completions API, and Batch API. It’s accessible to Free, Plus, and Team users on ChatGPT starting immediately. Enterprise users will get access next week.
OpenAI plans to add fine-tuning capabilities for GPT-4o mini soon. This will allow users to customize the model for specific tasks or domains.
The introduction of GPT-4o mini is part of OpenAI’s goal to make AI more affordable and accessible. They aim to continue reducing costs while improving AI capabilities. This new model is a significant step towards integrating powerful AI into everyday applications.
GPT-4o mini makes advanced intelligence more accessible to developers and users alike. It opens up new possibilities for AI integration in various fields and applications.
