---
title: 'RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control'
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/2307-15818-rt-2.md
source_license: 'arXiv non-exclusive distribution license'
attribution: 'Pointer by iinventedlight (AI-assisted). Source: Anthony Brohan and 53 coauthors (listed on the arXiv abs page), RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control (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: AI-for-science'
tags: arxiv, AI-for-science, robotics, vision-language-action
arxiv_id: 2307.15818
alias: 2307.15818-rt-2
type: pointer
format: pointer
sources:
  - id: S1
    title: 'RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control'
    publisher: arXiv
    authors: 'Anthony Brohan and 53 coauthors (listed on the arXiv abs page)'
    url: https://arxiv.org/abs/2307.15818
    version_read: 'v1 (28 Jul 2023)'
    accessed: 2026-11-12
    license: 'arXiv non-exclusive distribution license'
---
# RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control

## Summary
[RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control](https://arxiv.org/abs/2307.15818) (2023) introduced RT-2, a vision-language-action model that transfers web-scale vision-language knowledge into robot control. The authors train a single end-to-end model that maps robot observations to actions while still benefiting from large-scale pretraining on language and vision-language data from the web.

What it introduced: a way to treat robot actions as a language-like output of a vision-language model, so that web knowledge and semantic reasoning can show up in control, instead of training a robot policy only on limited demonstration data.

What it led to: later open vision-language-action models and multi-robot datasets. Live notes: [2310-08864-open-x-embodiment](https://iinventedlight.com/library/notes/2310-08864-open-x-embodiment/). This week's drafts also include [2406-09246-openvla](https://iinventedlight.com/library/notes/2406-09246-openvla/).

Use it when you want the closed RT-2 result alongside later open follow-through. Where to read it: https://arxiv.org/abs/2307.15818 (v1 (28 Jul 2023)).

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).
