As competition grows in the field of generative neural networks, OpenAI is strengthening its position by introducing updated versions of language models and reducing the cost of their use. Here are the main innovations:
- Software Feature Description Tool: OpenAI has introduced an innovative tool that allows developers to describe software features in natural language and automatically generate the appropriate code to execute them. This opens up opportunities for creating chatbots that can process queries, interact with databases, and extract structured data from text.
- Expanded Context Window: The GPT-3.5-turbo model now has a wider context window that can accommodate up to 16k tokens. This increase helps the model better understand the meaning of requests and generate more relevant responses without deviating from the topic. Currently, OpenAI is also testing the GPT-4 model with a context window of 32k tokens.
- Reduced cost of use: The cost of the standard version of GPT-3.5-turbo without extended context window has been reduced by 25%, now $0.0015 per 1000 input tokens and $0.002 per 1000 output tokens. This means that for $1 you can get about 700 pages of text. Also, the cost of using the text-embedding-ada-002 model, designed to measure text similarity, has been reduced by 75% to $0.0001 per 1,000 tokens.
With updated language models and a reduced cost of use, OpenAI simplifies and expands access to powerful text generation tools, opening up new perspectives for developers and content creators, cocnludes NIXSolutions.