OpenAI 12-Day Conference – Day 9
On the 9th day of OpenAI’s 12-day release event, several exciting updates were introduced that promise to reshape how developers and businesses use AI models. Let’s dive into the key highlights:
1. Strong Reasoning Model “O1” Now Supports API Calls and Image Recognition
Previously, OpenAI’s O1 model was available only through the web interface. Today’s release brings significant improvements, as O1 now supports API calls, allowing businesses and developers to integrate it into their applications. Additionally, O1 can now also recognize images, expanding its capabilities beyond just text-based input.
Quick Comment: While this feature might seem expected, API access was something OpenAI had to offer sooner or later. Using only a web interface for large-scale usage wasn’t sustainable.
2. 60% Price Drop for Lightweight Model “4O” Speech Recognition
OpenAI has reduced the price of its 4O model’s speech recognition feature by 60%, now charging just $40 per million tokens.
Quick Comment: While it is still more expensive than some competitors, the price drop is certainly a good move. The 4O model still delivers more natural and fluent responses compared to other models, justifying its slightly higher cost.
3. Introduction of the “4O Mini” at a 90% Discount
A major highlight today was the introduction of the 4O Mini, a smaller and more affordable version of the 4O model. Priced at just $10 per million tokens, it represents a 90% price reduction compared to the original model. The 4O Mini also offers faster response times.
Quick Comment: The 4O Mini is a game-changer for AI applications. Although the 4O model was already quite lightweight, many use cases don’t need such a large model. The 4O Mini is ideal for portable devices like smartphones, AR glasses, and smartwatches. It also opens up the possibility of integrating AI into toys, creating interactive AI-powered toys that can engage with children.
Conclusion
Today’s announcements from OpenAI mark a big leap forward in both pricing and capability for AI models. The O1 model’s ability to support API calls and image recognition opens up a wide range of possibilities for developers. Meanwhile, the price drop for the 4O model and the launch of the 4O Mini make AI technology more accessible for a variety of applications. These changes are expected to lower costs and improve the user experience in everyday devices, making AI more practical for things like voice interaction on mobile devices and even AI-driven toys.