OpenAI’s Most Advanced GPT Model Now Comes With A ‘Mini’ Version
Image via OpenAI
OpenAI, the research lab behind ChatGPT and DALL-E 3, has unveiled a new addition to its Generative Pre-trained Transformer (GPT) family—the GPT-4o mini, its “most cost-efficient small model.” This smaller, more streamlined version of the powerful GPT-4o promises to free up more access to advanced AI capabilities at lower rates and faster processing times.
GPT-4o mini excels in text-based tasks, outshining models like GPT-3.5 Turbo in both textual intelligence and multimodal reasoning, as well as excelling on benchmarks like MGSM (mathematical reasoning) and HumanEval (coding). With a 87% score on both benchmarks, GPT-4o mini demonstrates a prowess in tackling complex tasks. Furthermore, an expansive 128K token context window and support for a multitude of languages make it well-equipped for diverse projects.
Screenshot via OpenAI
Perhaps the most compelling aspect of GPT-4o mini is its budget-friendly pricing. At just 15 cents per million input tokens and 60 cents per million output tokens, it's over 60% cheaper than GPT-3.5 Turbo. This cost-efficiency empowers developers and businesses to integrate AI into their operations without exorbitant costs. This opens doors for innovation across industries, allowing for tasks ranging from customer service chatbots to intricate data analysis.
Introducing GPT-4o mini! It’s our most intelligent and affordable small model, available today in the API. GPT-4o mini is significantly smarter and cheaper than GPT-3.5 Turbo.https://t.co/sqJsFEYHWq pic.twitter.com/g6jMttp1mF
— OpenAI Developers (@OpenAIDevs) July 18, 2024
The GPT-4o mini’s capabilities extend beyond crafting email templates and essays. Its current API supports both text and vision, with plans to incorporate audio and video functionalities in the future.
Wild. GPT-4o on ChatGPT can generate full blown charts and statistical analysis from spreadsheets with a single prompt in less than 30 seconds.
— Zain Kahn (@heykahn) May 14, 2024
This stuff use to take ages in Excel. pic.twitter.com/Mvh5arWOQh
OpenAI has already begun integrating GPT-4o mini into their popular ChatGPT platform. Free, Plus, and Team users can now experience the benefits of this advanced model, which replaces the previously used GPT-3.5 Turbo. Enterprise users will have access shortly after.
“We envision a future where models become seamlessly integrated in every app and on every website,” explains the organization. “GPT-4o mini is paving the way for developers to build and scale powerful AI applications more efficiently and affordably.”
[via CNET and TechCrunch, images via OpenAI]
