---
title: "Mind2Web: Towards a Generalist Agent for the Web"
license: CC-BY-4.0
license_url: https://creativecommons.org/licenses/by/4.0/
license_note: 'paper is arXiv non-exclusive distribution license (verified on arXiv abs page 2026-11-12); this pointer note is our own words and is CC BY 4.0; we link to the paper rather than reproduce it'
source: https://iinventedlight.com/library/notes/2306-06070-mind2web.md
source_license: "arXiv non-exclusive distribution license"
attribution: 'Pointer by iinventedlight (AI-assisted). Source: Xiang Deng, Yu Gu, Boyuan Zheng, and 5 coauthors, Mind2Web: Towards a Generalist Agent for the Web (arXiv non-exclusive distribution license). This note links to the paper rather than reproducing it.'
author: 'iinventedlight (AI-assisted)'
date: 2026-11-12
status: published
approved_by: "Cole Withington, 2026-11-12 14:44 PT, Librarian chat widget t51s316: \"Yes, publish all 100 second-list arXiv notes\""
checked_on: 2026-11-12
as_of: 2026-11-12
confidence: medium
untested: 'Pointer note only. Paper not reproduced. Abs page and license checked 2026-11-12.'
category: arxiv
topic_area: 'ai-research: agents'
tags: arxiv, agents, web-agents, mind2web, benchmarks
arxiv_id: 2306.06070
alias: 2306-06070-mind2web
type: pointer
format: pointer
sources:
  - id: S1
    title: "Mind2Web: Towards a Generalist Agent for the Web"
    publisher: arXiv
    authors: "Xiang Deng and 7 coauthors (listed on the arXiv abs page)"
    url: https://arxiv.org/abs/2306.06070
    version_read: "v3 (9 Dec 2023)"
    accessed: 2026-11-12
    license: "arXiv non-exclusive distribution license"
---
# Mind2Web: Towards a Generalist Agent for the Web

## Summary
[Mind2Web: Towards a Generalist Agent for the Web](https://arxiv.org/abs/2306.06070) (2023) is a dataset and benchmark for generalist web agents that follow language instructions on many real websites, not only on a simulated store or a small set of sites.

What it introduced: more than 2,000 open-ended tasks from 137 websites in 31 domains, with crowdsourced action sequences. The authors argue that a generalist web agent needs diverse domains, real websites rather than simplified simulators, and a wide range of interaction patterns. They also show that feeding raw HTML to a large model is often too expensive, and that a smaller model can first filter the page to make the large model more effective, including on sites or domains it has not seen.

What it led to: a standard real-website testbed that later multimodal and end-to-end web agents built on. Live notes include [2307-13854-webarena](https://iinventedlight.com/library/notes/2307-13854-webarena/). WebShop is a full note in this week's batch F; VisualWebArena and WebVoyager are pointer notes in this batch.

Where to read it: https://arxiv.org/abs/2306.06070 (v3, last revised 9 Dec 2023). NeurIPS 2023 Spotlight. Project page: https://osu-nlp-group.github.io/Mind2Web.

The paper is posted on arXiv under the arXiv non-exclusive distribution license. This note links to it rather than reproducing it. The note itself is CC BY 4.0.

License: this note is CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Pointer by iinventedlight (AI-assisted). Paper license: arXiv non-exclusive distribution; this note links to it and does not reproduce it.

---

Agents: [/agents.md](/agents.md) | [/.well-known/agent-library.json](/.well-known/agent-library.json).
