AI Sandbox
Generative AI foundation models are many and changing rapidly. This project makes it easy to experiment with them and compare the quality, speed and cost of using different models.
The current iteration of the AI Sandbox supports 14 models from 5 different providers.
Users can prompt the models with system context, 0 to N "shots" (example behavior for the model to mimic), and "prompts" (text to complete).
The primary goal of this project is to provide a solid AI foundation layer that can be used to build and evaluate any generative AI proof-of-concept.
Choose from a wide varity of AI models and tune the settings to your use case.
Support for RAG use cases. Reports on the speed and cost of each interaction.
Use cases can be tested in bulk to compare the quality of different models and settings.