OpenAI introduced an enhanced version of GPT-4 Turbo on Wednesday, making it accessible through an Application Programming Interface (API) and for subscribers of ChatGPT Plus, several media outlets reported. GPT-4 Turbo represents the company's latest generative AI model, now trained with data up to December 2023.
The updated model boasts new data processing capabilities, enabling analysis of text in JSON format and executing function calls to generate JSON code snippets. These snippets empower developers to automate tasks within their connected applications, such as sending emails, posting online, or making purchases, as reported by The Indian Express.
With a context window now expanded to accommodate up to 128,000 tokens, developers can leverage the new API capability to deploy GPT-4 Turbo for image analysis, including food recognition from various global cuisines. Additionally, the model demonstrates enhanced proficiency in generating computer code. Developers with access to the GPT-4 Turbo API can seamlessly integrate the latest model into their applications and services, the national-daily said.
The upgraded GPT-4 Turbo is now capable of generating fully functional websites by producing genuine code and interpreting image content and context to generate meaningful output.
The version of GPT-4 Turbo featuring vision capabilities through JSON mode is now accessible to ChatGPT Plus subscribers at a cost of $20 per month. This version is also available through popular editions of ChatGPT on the ChatGPT store. Users can update to ChatGPT-4 Turbo through Microsoft's Copilot, choosing either creative or precise mode.
Earlier this year, in January, OpenAI, under the leadership of Sam Altman, introduced several updates to its AI models and pricing structure to address escalating competition. The pricing for GPT-3.5 Turbo was reduced, and GPT-4 Turbo received a performance boost following user complaints regarding the model's perceived "laziness."
While the industry eagerly anticipates the release of GPT-5, other models have made significant strides in matching or surpassing the capabilities of their predecessors. Anthropic's Claude 3 Opus, released in March, has surpassed GPT-4 in the closed-source models leaderboard, closely followed by Google's Gemini Pro, as reported by IE.