Open source models have several advantages in the business applications. They are becoming increasingly popular due to their ability to create custom training sets using outputs from larger, more powerful models. This allows smaller models to achieve similar performance to the larger models. Open source models also have the advantage of being able to stack different fine tunings, allowing them to perform multiple tasks simultaneously. This enables quick experimentation and makes it difficult for larger centralized companies to compete. While there are advantages to proprietary models, such as high quality, larger input window sizes, and faster speed, open source models can still be sufficient for many use cases that don't require a super powerful model. By starting with a general purpose model and making it specific to a particular task, open source models can serve specific needs effectively.