I've found that for the most part the articles that I want summarized are those which only fit the largest context models such as Claude. Because otherwise I can skim-read the article possibly in reader mode for legibility.

Is llama 2 a good fit considering its small context window?