The AI research organization introduces a suite of updates to improve developer experience and reduce costs in UK.
OpenAI has announced a comprehensive update to its suite of AI models and API tools, including the launch of advanced embedding models, an updated GPT-4 Turbo, enhanced moderation models, and new API usage management tools.
In addition, the organization is set to reduce the pricing on GPT-3.5 Turbo to further support its developer community.
>>> OpenAI: New embedding models and API updates
OpenAI is rolling out two new embedding models: the compact text-embedding-3-small and the more robust text-embedding-3-large. These models are designed to convert content into numerical sequences, simplifying the process for machine learning models to perform tasks like clustering or retrieval.
The text-embedding-3-small model offers a significant performance boost over its predecessor, while the text-embedding-3-large model sets a new standard for OpenAI's embedding performance.
Additionally, the cost for the smaller model has seen a dramatic reduction, making it even more accessible for developers.
text-embedding-3-small
|
text-embedding-3-large
|
|
Price
|
$0.00002/1k tokens
|
$0.00013 / 1k tokens
|
MIRACL average
|
44.0 | 54.9 |
MTEB average
|
62.3 | 64.6 |
The upcoming GPT-3.5 Turbo model will not only be more affordable but will also include improvements in accuracy and bug fixes for non-English language functions.
For pricing, the new model features a 50% reduction in input costs to $0.0005/1k tokens and a 25% decrease in output costs to $0.0015/1k tokens.
It also offers enhanced accuracy for format-specific responses and resolves a bug that affected text encoding in non-English language function calls.
Developers using the pinned GPT-3.5 Turbo model alias will be automatically upgraded to the new version, benefiting from the reduced input and output prices.
The updated GPT-4 Turbo preview model promises more thorough task completion, addressing previous performance concerns.
Since its launch, over 70% of GPT-4 API customers have switched to GPT-4 Turbo due to its more recent knowledge base, expanded 128k token context windows, and lower pricing.
The new update introduces gpt-4-0125-preview, which improves task completion like code generation and addresses the issue of "laziness" where previous models might not finish a given task. It also includes a fix for non-English UTF-8 text generation bugs.
Additionally, OpenAI has created an alias model name, gpt-4-turbo-preview, that will automatically update users to the latest version of GPT-4 Turbo preview as it's released.
The full release of GPT-4 Turbo with visual capabilities is scheduled for the near future.
>>> AI App Promotion Solutions
The new text-moderation-007 model is part of OpenAI's dedication to building safe AI systems. It will serve as the most robust moderation model to date, helping developers to easily identify potentially harmful text.
OpenAI is rolling out two platform enhancements aimed at providing developers with greater insights into their API usage and more granular control over their API keys.
Developers can now set specific permissions for each API key directly from the API keys page, allowing for tailored access like read-only or endpoint-specific restrictions.
Moreover, after enabling tracking, the updated usage dashboard and export feature now offer detailed metrics broken down by individual API keys.
This facilitates easy monitoring of usage patterns across different features, teams, products, or projects when distinct API keys are used for each category.
OpenAI has plans to further improve API management capabilities and is preparing to launch GPT-4 Turbo with vision in general availability. The organization encourages developers to follow its updates and engage with the community on social media platforms.
Click "Learn More" to drive your apps & games business with the ASO World app promotion service now.
These updates from OpenAI represent a significant step forward in the democratization of AI technology. By enhancing performance, reducing costs, and offering new tools for safe and efficient API management, OpenAI is ensuring that developers at all levels can access state-of-the-art AI capabilities.
The focus on embedding models, in particular, underscores the importance of machine learning applications that require nuanced understanding and processing of large datasets.
Get FREE Optimization Consultation
Let's Grow Your App & Get Massive Traffic!
All content, layout and frame code of all ASOWorld blog sections belong to the original content and technical team, all reproduction and references need to indicate the source and link in the obvious position, otherwise legal responsibility will be pursued.