Curie vs AI-Scientist-v2
AI agents for scientific experiments compared
Curie and AI-Scientist-v2 use AI agents to plan and run scientific experiments, analyze results, and support research. Curie emphasizes reproducible workflows using a researcher's data or starter code, while AI-Scientist-v2 explores open-ended machine-learning ideas with tree search and can draft papers.
Curie: Automate Scientific Experiments with AI Agents
Curie is a Python framework that uses AI agents to plan, run, analyze, and report on scientific experiments. It suits researchers who want to automate exploratory or machine-learning work, but its inactive development and lack of recent releases warrant caution.
AI-Scientist-v2: Automate Scientific Experiments and Paper Drafts
AI-Scientist-v2 uses LLM agents and tree search to propose research ideas, run experiments, analyze results, and draft papers. It is aimed at researchers exploring open-ended machine learning questions who can provide model APIs and a controlled compute environment.
| Curie | AI-Scientist-v2 | |
|---|---|---|
| Language | Python | Python |
| License | Apache-2.0 | NOASSERTION |
| Stars | 370 | 7.3k |
| Forks | 32 | 995 |
| Last analyzed | Oct 9, 2026 | Oct 3, 2026 |
Key differences
- Curie supports machine-learning engineering, system analysis, and broader scientific discovery; AI-Scientist-v2 focuses on open-ended machine-learning research.
- Curie accepts user-provided datasets and starter code and includes verification modules; AI-Scientist-v2 generates and refines ideas, then explores alternatives with best-first tree search.
- Curie produces experiment reports and records artifacts; AI-Scientist-v2 also produces manuscript drafts and supports citation and review stages.
- Curie lists an Apache-2.0 license, while AI-Scientist-v2 lists NOASSERTION and describes a project-specific license with a mandatory disclosure requirement for resulting manuscripts.
- Curie requires Docker and configured model credentials; AI-Scientist-v2 documents a Linux, NVIDIA GPU, CUDA, and PyTorch setup, plus supported model APIs.
- Curie has 370 stars and 32 forks; AI-Scientist-v2 has 7.3k stars and 995 forks. These counts alone do not establish suitability or reliability.
Choose Curie if you…
- need an agent to investigate a question using your own dataset or starter code.
- want experiment reports, recorded artifacts, and verification modules aimed at reproducible procedures.
- want support for system analysis or scientific discovery beyond machine-learning research.
Choose AI-Scientist-v2 if you…
- want to explore open-ended machine-learning ideas through tree search rather than use fixed experiment templates.
- need an exploratory pipeline that can draft a manuscript and support citation and review stages.
- can provide a controlled compute environment and review uncertain results and generated claims.
Compare the numbers side by side, or add more projects
This comparison is generated with AI from the OSRepos analyses of both projects. Always check each project's repository and documentation before choosing.