OpenAI, the artificial intelligence research company, has introduced a new generation of embedding models. These models can convert text into a numerical form that can be used for various machine learning tasks. Along with the embedding models, the company has also released new versions of its GPT-4 Turbo and moderation models, as well as new API usage management tools.
New Embedding Models
The new embedding models, text-embedding-3-small and text-embedding-3-large, offer stronger performance and reduced pricing compared to the previous generation model, text-embedding-ada-002. These models are capable of creating embeddings with up to 3072 dimensions, which captures more semantic information and improves the accuracy of downstream tasks. OpenAI claims that the average scores on commonly used benchmarks for multi-language retrieval (MIRACL) and English tasks (MTEB) have significantly improved with the new models.
“The new embedding models provide stronger performance and reduced pricing, making them more accessible for developers to use in their machine learning tasks.” – OpenAI
Updates to GPT-4 Turbo and Moderation Models
The company has also updated its GPT-4 Turbo and GPT-3.5 Turbo models, which are large multimodal models capable of understanding and generating natural language or code. The new versions of these models come with improved instruction following, JSON mode, more reproducible outputs, and parallel function calling. In addition, OpenAI has introduced a new 16k context version of GPT-3.5 Turbo, allowing for the processing of longer inputs and outputs compared to the standard 4k version.
Furthermore, OpenAI has enhanced its text moderation model, improving its capability to detect sensitive or unsafe text. The updated version now supports more languages and domains and provides explanations for its predictions.
Improved API Usage and Pricing
OpenAI has introduced new ways for developers to manage API keys and understand API usage. Developers can now create multiple API keys with different permissions and scopes, and monitor their usage and billing details on the OpenAI Dashboard. Additionally, the company has reduced the pricing of the GPT-3.5 Turbo model by 25%, making it more accessible for developers to build applications.
“We are continuously working to improve our models and services, and to make them more useful and affordable for developers and customers.” – OpenAI
The company encourages developers to contribute evaluations to help improve the models for different use cases. OpenAI is committed to releasing new models, features, and tools in the future, further expanding the applications and capabilities of artificial intelligence.
