---
format: "aidr-story-markdown/v1"
id: "c930db4a94e6e04430faa7c93c5f9fad3d046f0efe9f901d26326af7a2020bc0"
canonical_url: "https://aidr.today/c930db4a?lang=en"
title: "DoGBench: The first user-facing docs generation benchmark. No model scores >50%"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-10-01T23:10:32.000Z"
category: "Research"
topics: ["agent"]
source_urls: ["https://dogbench.ai/","https://news.ycombinator.com/item?id=49928118"]
summary: "DoGBench evaluates whether agents identify when user-facing software documentation needs updating and produce useful patches, using rubrics calibrated with project maintainers."
---

# DoGBench: The first user\-facing docs generation benchmark\. No model scores >50%

> [Open the canonical story](<https://aidr.today/c930db4a?lang=en>)

**Published:** 2026-10-01T23:10:32.000Z
**Category:** Research
**Topics:** agent

## Summary

DoGBench evaluates whether agents identify when user\-facing software documentation needs updating and produce useful patches, using rubrics calibrated with project maintainers\.

## Sources

- [Story source](<https://dogbench.ai/>)
- [Discussion](<https://news.ycombinator.com/item?id=49928118>)