Fine-tuning now available for GPT-4o

OpenAI has added fine-tuning for GPT-4o and GPT-4o mini, giving paid developers more control over model behavior.

The company is offering promotional free training tokens and has published pricing details for the customization feature.

Partner benchmarks from Cosine and Distyl show stronger results on software engineering and text-to-SQL tasks.

Early results suggest fine-tuning can meaningfully improve accuracy for specialized coding and SQL workloads.