idea: cortex runs with model alias

[louis-menlo]

Problem
I like using Cortex to run LM models on different engines, and there are also remote engines that boost my daily work. The problem is that I have to reconfigure the models from 3rd applications, that's quite tough.

I'm using Llama 3.1 as an alternative to Copilot with my IDE setup, but sometimes I want to switch to remote endpoints like Claude 3.5 Sonnet, for example. I have to go to my IDE plugin to update the model parameter, which is very unproductive.

What I want is that:
I can configure my IDE plugin with the model name copilot, then I go to Cortex and run cortex run llama3.1 copilot. Then, I can switch to another model without reconfiguring the IDE plugin by running cortex run claude-3-5-sonnet copilot, which would be very helpful.

Success Criteria
cortex run supports model aliases.

[louis-menlo]

It's outdated. It is supported on the latest version (via import)