OpenAI Introduces Fine-Tuning for GPT-4o

OpenAI has announced the introduction of fine-tuning features for its GPT-4o model, a highly anticipated update for developers. To make it even more appealing, OpenAI is offering one million complimentary training tokens daily for every organization until 23rd September.

Customizing GPT-4o with bespoke datasets can lead to better performance and cost savings for specific tasks. Fine-tuning allows precise control over the model's outputs, enabling customization of structure, tone, and the ability to adhere to detailed, domain-specific instructions.

Developers can achieve notable results using training datasets that contain as few as several dozen examples. This ease of access facilitates advancements in various fields, from tackling complex coding problems to refining creative writing.

"This is just the beginning," states OpenAI, emphasizing their dedication to continually expanding model customization options for developers.

GPT-4o fine-tuning is now available to all developers across all paid usage plans. Training costs are priced at 25 per million tokens, with inference costs set at 3.75 per million input tokens and $15 per million output tokens.

Additionally, OpenAI is offering GPT-4o mini fine-tuning with two million complimentary daily training tokens until 23rd September. To access this, choose "gpt-4o-mini-2024-07-18" from the base model dropdown on the fine-tuning dashboard.

The company has partnered with select organizations to test and explore the possibilities of GPT-4o fine-tuning:

OpenAI assures users that fine-tuned models remain fully under their control, with full ownership and confidentiality of all business data. This ensures no data sharing or utilization for training other models.

Strict safety protocols have been put in place to prevent the misuse of fine-tuned models. Continuous automated safety assessments are performed, along with usage monitoring, to ensure compliance with OpenAI's stringent usage policies.