The GPT-4 API is now open to everyone as OpenAI goes all-in on chat

The global tech industry is witnessing a milestone as OpenAI's highly anticipated GPT-4 API is widely available to all paying customers.

Since its launch in March, the GPT-4 API has been favored by developers around the world, but so far, it has only been available to some customers. Now, those developers will have the opportunity to take advantage of the capabilities of GPT-4, currently hailed as the most proficient model available.

GPT-4 for everyone

The GPT-4 API provides an 8K context window, which refers to the amount of textual information a model can "consider" or "remember" when generating a response. This 8K context window essentially means that the model considers the last 8,000 tokens (roughly 8,000 words or characters, depending on the language) to generate its output, a feature that is critical to keeping the model's responses consistent and coherent.

Existing API developers with a history of successful payments can access the GPT-4 API. OpenAI plans to expand access to new developers by the end of the month, and then increase the rate limit based on compute availability.

Additional API Versions and Development

The company also generally provides GPT-3.5 Turbo, DALL·E (image generation), and Whisper (audio) APIs, suggesting that the models are ready for production-scale use. Additionally, OpenAI is working on fine-tuning for GPT-4 and GPT-3.5 Turbo, which developers can look forward to later this year.

In the context of AI models, fine-tuning refers to taking a pre-trained model (one that has learned general patterns from large datasets) and customizing or "tuning" it to a more specific task or dataset.

This approach allows developers to leverage the extensive learning of the underlying model while tailoring its behavior to better suit their unique needs, increasing the accuracy and efficiency of their specific applications. This much-anticipated feature is something developers can look forward to throughout the year.

Chat complete

The rise of chat-based models used in GPT-4 is significant. Since its launch in March, the Chat Completions API now accounts for 97% of OpenAI's GPT usage, effectively replacing the old Completions API based on free-form text prompts. The shift to a more structured chat-based interface has proven game-changing, delivering more flexibility, specificity, and superior results.

However, these improvements did not come without a price. OpenAI announced the deprecation plan for the legacy Completions API model.

Effective January 4, 2024, the old completion model will be retired and replaced with a new one, as part of OpenAI's efforts to invest more in the Chat Completions API and optimize its computing power.

"While this API is still accessible, starting today we will mark it as 'legacy' in the developer documentation."

Developers who wish to continue using their fine-tuned models after January 4, 2024 will need to fine-tune replacement models on top of the new base GPT-3 model or updated models such as gpt-3.5-turbo and gpt-4.

Embedded models are deprecated

OpenAI also said that following these developments, the old "embedding" model will be deprecated. Users must migrate to "text-embedding-ada-002" by January 4, 2024. OpenAI has assured developers using older models that they will bear the financial cost of re-embedding content using these new models.

As OpenAI leads this dramatic shift, it also raises questions about the future of older models and the implications for the developers and companies that rely on them. This historic turn in the development of artificial intelligence underscores the rapid and relentless pace of technological innovation, shaping the future of industries across the globe.