OpenAI has announced the launch of a fine-tuning feature for its most advanced model, GPT-4o being a significant advancement for artificial intelligence customization. This new offering, unveiled on Tuesday, allows corporate customers and developers to tailor the AI model using their own data for just $25, starting from September 24, 2024.
According to a blog post by the company, the fine-tuning feature is now fully operational. It is designed to enable businesses and developers to adjust the model’s responses to meet specific requirements, such as modifying the structure and tone or adhering to complex, domain-specific instructions. This customization is available exclusively for GPT-4o, providing users with a highly adaptable AI tool.
This launch marks a major advancement in AI customization, fulfilling one of the most requested features by developers and further solidifying OpenAI’s leadership in the AI space.
Olivier Godement, OpenAI’s head of product for its API, emphasized the company’s dedication to making the fine-tuning process more accessible. “We’ve been extremely focused on lowering the bar, the friction, the amount of work it takes to get started,” Godement stated, underscoring OpenAI’s commitment to simplifying AI customization for its users.
The fine-tuning feature is available to all developers and companies on paid plans. To initiate the process, users can access the fine-tuning dashboard, select the base model, and begin customizing GPT-4o according to their specific needs.
In a bid to encourage widespread adoption, OpenAI is offering 1 million free training tokens per day for each organization until September 23, 2024. After this period, training will be priced at $25 per million tokens, with inference costs set at $3.75 per million input tokens and $15 per million output tokens. For those seeking a more budget-friendly option, GPT-4o mini fine-tuning is also available, with 2 million free training tokens per day during the same promotional period.
Key features of the GPT-4o fine-tuning include enhanced model performance in handling structure, tone, and complex instructions, leading to improved outcomes in specialized applications. The customization process is streamlined, typically requiring only one to two hours, allowing for rapid deployment of tailored solutions. OpenAI has also implemented robust safety protocols, including automated evaluations, to prevent misuse of fine-tuned models.
This new capability opens up vast possibilities for businesses. For example, a company could fine-tune GPT-4o to create a specialized customer-service chatbot capable of answering detailed product questions with minimal training data, making it an effective and accessible tool for a wide range of use cases.
OpenAI has assured users that fine-tuned models will remain fully under their control, with complete ownership of business data, including all inputs and outputs. The company has also implemented multiple layers of safety measures to ensure that fine-tuned models are not misused, including continuous automated safety evaluations and monitoring to ensure compliance with usage policies.