OpenAI Unleashes GPT-4o Fine-Tuning With Free Daily Tokens
OpenAI now allows developers to fine-tune its powerful GPT-4o model to better suit specific application needs. The company offers one million free tokens per day for this process through September 23.
OpenAI announces that developers can now fine-tune its flagship GPT-4o model to customize the artificial intelligence for specific organizational needs. This process allows users to adjust the model's tone, follow strict instructions, and improve accuracy in technical tasks using surprisingly small datasets. The capability builds upon the recent release of fine-tuning for the smaller and more affordable GPT-4o mini variant.
To encourage adoption of this new feature, OpenAI provides up to one million free tokens per day to any organization through September 23, 2024. Tokens serve as the numerical building blocks that AI models use to process and generate language, making them essential for training customized versions of the software. Developers access this tool directly through the OpenAI fine-tuning dashboard by selecting the specific GPT-4o base model.
While the initial promotional period offers significant savings, standard pricing takes effect after the free token allotment expires. OpenAI typically charges $25 per one million tokens for the fine-tuning process itself, with additional fees of $3.75 per million input tokens and $15 per million output tokens to run the customized model in production. Company technical staff emphasize that this expanded customization represents just the beginning of their ongoing investment in developer tools.