GPT-4o mini is OpenAI's latest small-scale AI model designed to make advanced artificial intelligence more accessible and affordable.
Cost-efficiency: It's priced at 15 cents per million input tokens and 60 cents per million output tokens, making it over 60% cheaper than GPT-3.5 Turbo and significantly more affordable than previous cutting-edge models.
Performance: GPT-4o mini outperforms GPT-3.5 Turbo and other small models on academic benchmarks for both textual intelligence and multimodal reasoning. It scores 82% on MMLU (Massive Multitask Language Understanding) benchmark.
Capabilities: It supports text and vision inputs, with plans to include video and audio capabilities in the future. The model can handle tasks such as chaining multiple API calls, processing large volumes of context, and interacting with customers through real-time text responses.
Context window: GPT-4o mini has a large context window of 128K tokens, allowing it to process extensive data like full codebases or conversation histories.
Knowledge cutoff: The model has knowledge up to October 2023.
Availability: GPT-4o mini is available through the OpenAI API for developers and in ChatGPT for Free, Plus, and Team users. Enterprise users will have access starting next week.
Applications: It can be used for various tasks including prototyping and design, user feedback analysis, content creation, code generation and analysis, data extraction and structuring, and multimodal reasoning.
Safety: GPT-4o mini incorporates the same safety mitigations as GPT-4o, with OpenAI continuously monitoring its use to ensure user safety and protection.
OpenAI envisions GPT-4o mini as a step towards making AI more accessible and integrated into everyday applications, paving the way for developers to build and scale powerful AI applications more efficiently and affordably.