The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Search
Summary
The AI Scientist-v2 is an end-to-end agentic system that iteratively formulates scientific hypotheses, designs and executes experiments, analyzes and visualizes data, and autonomously authors scientific manuscripts. Compared with v1, it drops human-authored code templates, generalizes across machine-learning domains, and uses a progressive agentic tree search run by an experiment-manager agent. A vision-language model also reviews figures. The authors submitted three fully autonomous manuscripts to a peer-reviewed ICLR workshop; one scored above the average human acceptance threshold, which they present as the first fully AI-generated paper to pass that workshop review. The authors have open-sourced the code at https://github.com/SakanaAI/AI-Scientist-v2. The paper also discusses the role of AI in science, including safety. [S1]
Claims
- id: C1 claim: "AI Scientist-v2 is an end-to-end system that hypothesizes, experiments, analyzes, and writes manuscripts without human code templates." source: S1 evidence: "eliminates the reliance on human-authored code templates"
- id: C2 claim: "One of three workshop submissions scored above the average human acceptance threshold." source: S1 evidence: "one manuscript achieved high enough scores to exceed the average human acceptance threshold"
- id: C3 claim: "Code is open-sourced on GitHub." source: S1 evidence: "https://github.com/SakanaAI/AI-Scientist-v2"
Verification
- Code available: Paper: https://github.com/SakanaAI/AI-Scientist-v2, HTTP 200 on 2026-11-12. Not run.
- Reproduced: not reproduced (2026-11-12). Code not run.
- Related work: The earlier system is in 2408-06292-the-ai-scientist. Tree search is discussed in 2305-10601-tree-of-thoughts.
- Limits: The acceptance result is for one workshop, not a journal or conference main track. The system is demonstrated on machine-learning experiments. Automated discovery still needs human judgment about validity and novelty.
- License: CC BY 4.0 on the abs page (2026-11-12; icon by-4.0.png; http://creativecommons.org/licenses/by/4.0/). Submitted 10 Apr 2025 (v1). Sources: [S1] https://arxiv.org/abs/2504.08066, v1 (10 Apr 2025). checked_on: 2026-11-12.
When to apply
Use when asking what changed between AI Scientist v1 and v2, or what the authors reported about a workshop submission.
Facts
- id: F1 claim: "AI Scientist-v2 runs hypothesis formation, experiments, analysis, and manuscript writing with a progressive agentic tree search." confidence: medium date: 2026-11-12 sources: [S1] license: CC BY 4.0
- id: F2 claim: "The authors report that one of three ICLR workshop submissions exceeded the average human acceptance threshold." confidence: medium date: 2026-11-12 sources: [S1] license: CC BY 4.0
- id: F3 claim: "Code is at https://github.com/SakanaAI/AI-Scientist-v2." confidence: medium date: 2026-11-12 sources: [S1] license: CC BY 4.0
Sources
- [S1] arXiv 2504.08066, https://arxiv.org/abs/2504.08066. License: CC BY 4.0. v1 (10 Apr 2025). Accessed 2026-11-12.
License: this note is CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Summary by iinventedlight (AI-assisted). Paper is CC BY 4.0.
License: CC-BY-4.0 · Summary by iinventedlight (AI-assisted). Source: Yutaro Yamada, Robert Tjarko Lange, Cong Lu, and 5 coauthors, The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Search (CC BY 4.0).
Cite this
'iinventedlight (AI-assisted)'. 2026. "The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Search." iinventedlight Library, iil:note:5b2b59sf, version of 2026-11-12. https://iinventedlight.com/library/notes/2504-08066-ai-scientist-v2/. License: CC-BY-4.0.
@misc{iil_note_5b2b59sf,
author = {{'iinventedlight (AI-assisted)'}},
title = {{The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Search}},
howpublished = {iinventedlight Library},
year = {2026},
date = {2026-11-12},
url = {https://iinventedlight.com/library/notes/2504-08066-ai-scientist-v2/},
note = {Permanent ID iil:note:5b2b59sf. License: CC-BY-4.0.},
}
Spotted a mistake? Use the MCP feedback tool or email hello@iinventedlight.com.
For language models and agents: start at /llms.txt | /agents.md | /.well-known/agent-library.json. This page in Markdown: /library/notes/2504-08066-ai-scientist-v2.md.