OpenAI recently held its first developer conference, unveiling significant news, including the launch of GPT-4 Turbo. This new AI system boasts two key characteristics: enhanced power and more cost-effective access. Additionally, there are noteworthy updates for GPT-4.
AI Assistants on the Rise
In recent months, AI assistants like GPT-4 have become the latest technological trend. Chances are, you’ve already explored their capabilities to some extent. Now, OpenAI, the creative force behind these advancements, has introduced its new model, which is expected to set the stage for the proliferation of many other AI systems.
GPT-4 Turbo: A More Effective AI
In the presentation of GPT-4 Turbo, one of its most significant features is its updated knowledge base. It now encompasses all events up to April 2023, offering greater flexibility and more up-to-date responses, aligning with real-world developments.
However, the increase in knowledge is just one aspect of this new version. GPT-4 Turbo can now handle prompts with a maximum of 128,000 tokens. To put this into perspective, this is enough text to fill around 300 pages, equivalent to a book like “Harry Potter and the Prisoner of Azkaban.”
Two Versions and Pricing
Developers working with the paid version of GPT-4 can access GPT-4 Turbo by utilizing the gpt-4-1106-preview update. OpenAI is expected to make it globally available in the coming weeks, following testing and observing its usage among early adopters.
GPT-4 Turbo is available in two versions:
- Text-Only: This version is suitable for classic chatbot interactions, accommodating text requests. It is available in the API update mentioned earlier.
- Image Analysis: The second version supports image analysis, enhancing its capabilities.
Pricing depends on the number of tokens used. Tokens serve as the unit of measurement for text input and responses within ChatGPT. They aren’t equivalent to letters or words but are used to gauge the text you input and the system’s responses. The price for GPT-4 Turbo is $0.01 for 1,000 tokens in the command request, and $0.03 for the same amount of tokens in the program’s responses. To give you an approximate idea, 1,000 tokens translate to around 700-750 words. OpenAI aims to make AI usage more cost-effective compared to GPT-4, opening up opportunities for tasks like text summarization. Pricing for images varies based on size, with the example of a 1080 x 1080-pixel image costing $0.00765.
GPT-4 Improvements
OpenAI’s commitment to AI excellence doesn’t stop with GPT-4 Turbo. The company has also announced improvements to GPT-4. OpenAI is focused on optimizing the program’s performance and addressing areas where it falls short of human-level performance. This endeavor has required substantial human effort to ensure that GPT-4 delivers exceptional performance. In a bid to enhance user experience, OpenAI has doubled the limit of tokens per minute, even though token prices for GPT-4 remain unchanged. These developments aim to provide a more satisfying user experience.
Other new features announced during the conference include the compatibility of GPT-4 Turbo and GPT-3.5 Turbo with JSON mode and the availability of GPTs, customizable versions of the chatbot for companies and professional users. This system empowers users to create their versions of ChatGPT, both for private and commercial use, without the need for programming expertise, fostering rapid AI expansion based on OpenAI technology.
