# Chatio Benchmark

> Real-world LLM benchmarks evaluating AI models on practical, everyday tasks. Compare Anthropic, OpenAI, xAI, Google, and more.

Preferred host: https://www.chatio.dev

## Primary pages

- [Home](https://www.chatio.dev/): Product overview and real-world LLM benchmark positioning
- [Leaderboard](https://www.chatio.dev/leaderboard): Rankings across helpfulness, empathy, instruction following, creative writing, and comprehension
- [Blog](https://www.chatio.dev/blog): Essays on everyday AI assistants and how we read the board
- [RSS](https://www.chatio.dev/feed.xml): Blog feed for crawlers and readers

## Blog posts

- [The caveman tax: why the best agents write worse](https://www.chatio.dev/blog/the-caveman-tax): Labs optimized internal reasoning for token-efficient agents. That same compression bleeds into final prose, so creative writing got worse even as helpfulness climbed. GPT-4.5 is still the writing high-water mark.

## Notes for crawlers

- Index marketing pages, the leaderboard, model detail pages, and the blog.
- Do not index `/chat`, `/admin`, or `/api/*`.
- Canonical URLs always use https://www.chatio.dev (www).
- Sitemap: https://www.chatio.dev/sitemap.xml