OpenAI's latest advancement, GPT‑4o Mini, marks a pivotal moment in the evolution of accessible artificial intelligence. Aiming to democratize intelligence further, the model is priced at a mere 15 cents per million input tokens and 60 cents per million output tokens, setting a new precedent in cost-efficiency within the realm of AI technology. By outperforming GPT‑41 in chat preferences on the LMSYS leaderboard and achieving 82% on the MMLU benchmark, it stands as the most formidable small AI model to date.
With its broad capabilities of handling both text and vision inputs and promising support for video and audio in the future, GPT‑4o Mini is poised to revolutionize application development. Its superior performance in textual and multimodal reasoning benchmarks outshines previous iterations like GPT‑3.5 Turbo by over 60% in cost efficiency. The practicality of the model is enhanced by its 128K token context window, ideal for large datasets or history-context utilization.
GPT‑4o Mini's efficacy stretches across various use cases from mathematical reasoning, scoring an exceptional 87.0% on MGSM, to coding proficiency, where it achieves 87.2% on HumanEval. These metrics underscore OpenAI's mission to enhance the utility and scalability of AI applications.
For an in-depth exploration of GPT‑4o Mini's groundbreaking features and potential applications, delve into OpenAI's comprehensive blog coverage. Learn more: Discover the Future of AI with GPT‑4o Mini.
Source: https://openai.com/blog/gpt-4o-mini-advancing-cost-efficient-intelligence
Top comments (0)