Out of curiosity - do you have an estimate on how much you've spent on various LLM API services through all of your experimentation?
And in your experience, what service do you feel hits a good sweet spot for performance/price if summarizing long text excerpts is the main use case? Inference time isn't an issue, this will be an ongoing background task.
And in your experience, what service do you feel hits a good sweet spot for performance/price if summarizing long text excerpts is the main use case? Inference time isn't an issue, this will be an ongoing background task.