The proliferation of open AI models makes it hard to select the best one for specific mobile applications.
Build a benchmarking tool that runs multiple models on a device and compares their performance across metrics.
Mobile developers would pay to save time in model selection.
Start with a few popular models and basic performance metrics.
The risk is rapid obsolescence as new models emerge.