# SakanaAI/AI-Scientist-v2

The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Search

Repository: https://github.com/SakanaAI/AI-Scientist-v2
Canonical: https://ross.abutalabs.com/products/ai-scientist-v2
Language: Python
License: NOASSERTION
License Family: other
Last push: 2025-12-19T07:46:32+00:00

## Health v2 (maintenance only)
Score: 46/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 58, release rhythm 35, longevity 36
- inputs: {"age_days": 512, "days_push": 257, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 7050, forks 976 (observed 2026-08-28T04:09:55.506740+00:00)

## What it is
An end-to-end autonomous AI research system that generates hypotheses, runs machine learning experiments, analyzes data, and writes complete scientific manuscripts without human-authored templates. It employs progressive agentic tree search guided by an experiment manager agent and produced the first fully AI-written paper accepted through peer review at an ICLR workshop.

## Use cases
- automate scientific research end to end with AI
- generate and test machine learning research hypotheses automatically
- have an AI agent run experiments and write a paper
- use agentic tree search for open-ended scientific discovery
- automate the ML experiment and analysis pipeline
- produce draft scientific manuscripts from an AI research idea
- run autonomous ablation studies and data analysis

## When to choose
- You want fully open-ended ML research exploration without relying on predefined templates
- You need a complete pipeline from idea generation through experiments to a written manuscript
- You have a Linux machine with NVIDIA GPUs, LLM API keys, and a sandbox (e.g., Docker) for running LLM-written code

## When to avoid
- You have a well-defined task with a strong starting template and need high success rates - the v1 AI Scientist may perform better
- You cannot safely sandbox execution of LLM-generated code, which poses security and process risks
- You need automation for non-machine-learning scientific domains, which this system does not target
- You lack a Linux/CUDA environment or cannot afford extensive LLM API usage

## Facets
- artifact type: application
- maturity: active
- function: agent-framework, machine-learning, deep-learning, data-science, workflow-automation, llm-inference
- domain: artificial-intelligence, machine-learning, large-language-models, data-science
- platform: python
- tags: ai-scientist, automated-scientific-discovery, agentic-tree-search, autonomous-research, hypothesis-generation, paper-writing, llm-agents, experiment-automation, peer-review, ai-agents, automation, linux, gpu, docker

## Member repositories
- SakanaAI/AI-Scientist-v2 (main) score 46

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:55.506740+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:40:31.722710+00:00, confidence not recorded.
  - readme: https://github.com/SakanaAI/AI-Scientist-v2 (fetched 2026-08-28T04:09:55.506740+00:00, sha 1603de9d0c26)
- Data as of 2026-08-30T08:39:29.467469+00:00.