# NVIDIA paper finds frequent context compaction keeps LLM accuracy while shortening context

> Original title: Highly recommended paper from NVIDIA.

- Source: Elvis Saravia
- Published: 2026-11-04T12:25:01.000Z
- Bad Influence: https://badinfluence.news/items/xbzqgqwxghn115fhi0ftscz1x
- Original: https://x.com/omarsar0/status/2117956281910829083

## Summary

Elvis Saravia highlights an NVIDIA paper arguing that frequent context compaction lets frontier language models keep accuracy on long-horizon tasks while using short contexts. The post cites Claude Opus 4.5 scoring 70.2 on SWE-bench Verified with one action per context, against 70.0 with full history, and median peak context falling from 32K to 8K tokens. The paper also notes that later output tokens can cost up to 7x more than earlier ones on frontier open-weight models.

## Post

Highly recommended paper from NVIDIA.

(bookmark it)

There is a lot of debate about whether to use context compaction aggressively.

They show that with frequent compaction, frontier language models lose little to no accuracy on a broad range of long-horizon and long-input tasks while keeping their context short.

With one action per context, Claude Opus 4.5 scores 70.2 on SWE-bench Verified against 70.0 with the full history. Median peak context drops from 32K to 8K tokens.

The agent carries a short summary of its state between steps instead of the whole transcript. Simple compaction at the end of a conversation often does better than the full history, and Opus goes from 70.0 to 74.8 with it.

This also matters for cost. Later tokens cost more to generate than earlier ones, up to 7x on frontier open-weight models.

The pattern holds on Terminal-Bench, FRAMES, BFCLv3 and multilingual SWE-bench, with modest losses in some settings.

Before paying for million-token agent runs, it's worth testing aggressive compaction on your own tasks.

https://arxiv.org/abs/2611.05306

Chat with Paper: https://academy.dair.ai/papers/you-dont-need-long-context-2611.05306

![](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Fmedia%2FHWR-A3haMAA_4rw.jpg&mode=thumb&exp=1794009600&sig=daefdfbd01010817)