Generative AI juggernaut OpenAI has introduced new ways for developers to manage and track the use of API keys. The company also updated its GPT-3.5 Turbo and GPT-4 large language models, and unveiled two new text embedding models.
OpenAI made the rollout on January 25. Two API key managment platform improvements aim to give developers more visibility into usage and control over APIs. First, developers now can assign permissions to API keys from the API keys page. For example, a key could be assigned read-only access to power an internal tracking dashboard or restricted to access only certain endpoints.
Second, the API usage dashboard and export function now expose metrics for an API key after turning on track. This makes it simple to look at usage on a per team, feature, product, or project level by having separate API keys for each. More improvements are planned to improve the ability of developers to view API usage and manage keys, OpenAI said.
OpenAI unveiled two new text embedding models, a “smaller and highly efficient” text-embedding-3-small model and a “larger and more powerful” text-embedding-3-large model. The company touts text-embedding-3-large as its new best-performing embedding model.
An embedding is a sequence of numbers representing concepts within content such as code or natural language, OpenAI explained. Embeddings make it easy for machine learning models and other algorithms to understand the relationship between content and to perform tasks such as retrieval or clustering. These power applications such as knowledge retrieval in ChatGPT or the Assistants API.
The updated GPT-4 Turbo preview model, gpt-4-0125-preview, completes tasks such as code generation more thoroughly than the previous model and is intended to reduce “laziness” where the model does not complete a task, OpenAI said. The new model also includes a fix for a bug impacting non-English UTF-8 generations. For users who want to be automatically upgraded to new GPT-4 Turbo preview versions, OpenAI is introducing the gpt-4-turbo-preview model name alias, which will point to the latest GPT-4 Turbo preview model.
The company is also introducing a new GPT-3.5 Turbo model, the gpt-3.5-turbo-0125, and reducing prices on GPT-3.5 Turbo usage. The update improves the model’s accuracy at responding in requested formats and fixes a bug that caused a text encoding issue for non-English language function calls.
Copyright © 2024 IDG Communications, Inc.
