from Hacker News

Ask HN: What is the current sota model for abstractive text summarisation?

by meander_water on 9/1/25, 7:36 AM with 0 comments

I need to summarise a bunch of long form text, and I'd ideally like to run it locally.

I'm not an NLP expert, but from what I can tell, the best evaluation benchmarks are G-Eval, SummEval and SUPERT. But I can't find any recent evaluation results.

Has anyone here run evaluations on more recent models? And can you recommend a model?