# Zyphra's ZHA matches dense models on SWE-bench and beats them on LongBench v2

> Original title: We test ZHA on reasoning, coding, and long-context tasks: AIME 2026, SWE-bench Verified, Terminal-Bench 2.1, GPQA, RULER, and LongBench v2.

- Source: Zyphra
- Published: 2026-11-02T22:04:54.000Z
- Bad Influence: https://badinfluence.news/items/rcitfxmu1icnih3zoy9o7drgk
- Original: https://x.com/ZyphraAI/status/2117377435876655563

## Summary

Zyphra says its ZHA model is tested on AIME 2026, SWE-bench Verified, Terminal-Bench 2.1, GPQA, RULER, and LongBench v2. Deep ZHA matches a dense model on SWE-bench and outperforms it on LongBench v2 while attending to about 1.1% of decode keys.

## Post

We test ZHA on reasoning, coding, and long-context tasks: AIME 2026, SWE-bench Verified, Terminal-Bench 2.1, GPQA, RULER, and LongBench v2.

Deep ZHA matches dense on SWE-bench. On LongBench v2, it outperforms dense while attending to ~1.1% of decode keys.

![](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Fmedia%2FHWJsCK_boAAgxU4.png&mode=thumb&exp=1793923200&sig=30d345edb7f1b0ea)