Replace all API calls with mocked responses from actual API calls.
This reduces cost for API calls on every push and merge.
Also now affordable to matrix test on merge.
Also was unnecessary since only want to test the code and there's no practical way to test responses from LLM yet.
Don't need to test every LLM model since mocking responses.
This is also way faster - 17 seconds v. > 26 minutes when making API calls for every test and testing every model.
Replace all API calls with mocked responses from actual API calls. This reduces cost for API calls on every push and merge. Also now affordable to matrix test on merge. Also was unnecessary since only want to test the code and there's no practical way to test responses from LLM yet. Don't need to test every LLM model since mocking responses. This is also way faster - 17 seconds v. > 26 minutes when making API calls for every test and testing every model.