OpenAI has announced significant updates, including the launch of new embedding models, enhancements to GPT-4 Turbo, and the introduction of moderation models, alongside API usage management tools. Notably, the pricing for GPT-3.5 Turbo is set to decrease, making these advanced technologies more accessible.
The new embedding models, text-embedding-3-small and text-embedding-3-large, offer improved efficiency and performance, with the former being 5X cheaper than its predecessor. These models are designed to understand the relationships between content, aiding in tasks like clustering or retrieval, and can be adjusted in size without losing their concept-representing properties.
Additionally, the GPT-3.5 Turbo model will see a price reduction and improvements in accuracy and bug fixes. The GPT-4 Turbo preview has been updated to enhance task completion and address previous issues, with a plan to launch GPT-4 Turbo with vision in general availability soon.
The moderation model has also been updated to better identify potentially harmful text, contributing to safer AI systems. New platform improvements offer developers greater visibility and control over API usage and key management, promising further enhancements for larger organizations in the future.
These updates signify OpenAI’s commitment to making AI more efficient, accessible, and safe for developers and users alike.