OpenAI offers fine-tuning for GPT-4o users to customize AI model
- OpenAI now allows paying users of its GPT-4o and GPT-4o mini to fine-tune the artificial intelligence model for specific use cases, the Microsoft-backed (NASDAQ:MSFT) startup announced today.
- Fine-tuning is the process of taking a pre-trained large language model and adjusting it to improve the performance of the model and to better fit a user’s specific needs.
- For example, Cosine AI, a UK-based startup, developed Genie, an AI software engineering assistant that is powered by a fine-tuned GPT-4o model. It is now more adept at solving technical problems and can change coding with higher accuracy, while also using fewer tokens. It is considered the world’s most capable AI software engineering assistant.
- OpenAI said the fine-tuning feature has been one of the most highly requested features from developers.
- OpenAI is not the first to offer fine-tuning. Clarifai, Nebius AI, Cohere, Hugging Face, Amazon’s (AMZN) SageMaker, and Microsoft’s Azure AI Studio all allow for the fine-tuning of LLMs.