Fine-tuning now available for GPT-4o

AI Impact Summary

GPT-4o now supports fine-tuning, enabling customization on domain data to improve task-specific accuracy and alignment. This creates a new data-to-model pipeline (data collection, formatting, and safety reviews) and a training lifecycle that affects deployment timing, cost, and governance. Expect variability in performance across domain prompts and plan for ongoing evaluation, monitoring, and rollback strategies to maintain reliability.