AI-Scientist vs AI-Scientist-v2
AI-Scientist is much bigger: 14.5k stars against 7.1k. Over the days we have tracked them AI-Scientist moved +1.6% and AI-Scientist-v2 +2.7%, so AI-Scientist-v2 is growing faster right now.
AI-Scientist leads on model freedom. AI-Scientist-v2 does not take any axis by a clear margin.
Stars and commit dates come from our own daily tracking. The six axes are read off each project's documentation by our review pipeline, so they describe what a project says about itself, not what we measured in its code.
Where they stand today
Automates end-to-end scientific discovery by generating hypotheses, running experiments, and producing full LaTeX papers from templates.
- Stars
- 14.5k
- Tracked growth
- +1.6%
- Maturity
- ●●●●●
- Last commit
- 260d ago
- Language
- Jupyter Notebook
- License
- NOASSERTION
- Cost to run
- Requires your API keys for cloud models; local GPU needed for open-weight runs.
End-to-end autonomous scientific discovery: it uses progressive agentic tree search to ideate, run experiments, analyze results, and draft papers without human-authored templates.
- Stars
- 7.1k
- Tracked growth
- +2.7%
- Maturity
- ●●●●●
- Last commit
- 260d ago
- Language
- Python
- License
- NOASSERTION
- Cost to run
- Your API key, per-run (README cites ≈$15–$20 for experiments + ≈$5 for writing with default models)
Six axes, head to head
Each axis runs 0 to 5. The label under a score is what that project's own docs claim, not a category average.
| Axis | AI-Scientist | AI-Scientist-v2 |
|---|---|---|
Context depth How much of your codebase it sees before it answers: the open diff, the diff plus related files, or the whole repository. | ●●●●● Template-level view | ●●●●● Related files |
Noise control How it keeps output volume down — severity thresholds, deduplication, incremental runs over new commits only. | ●●●●● Ensemble reviews | ●●●●● Config-based controls |
Customization How far it bends to your team: custom rules, prompts, style guides, per-path config. | ●●●●● Template + prompt config | ●●●●● Configurable params |
Privacy Whether your code stays on your own infrastructure: fully local, self-hostable, or cloud API only. | ●●●●● Self-hostable (local GPU) | ●●●●● Cloud APIs (your keys) |
Model freedom Whether you can point it at any provider, or it is wired to one. | ●●●●● BYOK + local models | ●●●●● Multiple providers |
Setup ease What it takes to get a first useful run out of it. | ●●●●● GPU + dependencies | ●●●●● GPU install required |
Which one to pick
Pick AI-Scientist if…
End-to-end experiment automation — pick this when you want a research-focused system that generates ideas, runs experiments on local GPUs, and compiles full papers automatically.
- Model freedom: BYOK + local models (5/5 against 3/5)
Pick AI-Scientist-v2 if…
End-to-end autonomy — pick this when you want a research pipeline that generates hypotheses, executes experiments, and produces writeups using configurable agentic tree search across multiple LLM providers.
What people want from each one
SakanaAI/AI-Scientist
SakanaAI/AI-Scientist-v2
Questions people ask
Is AI-Scientist better than AI-Scientist-v2?
AI-Scientist leads on model freedom. AI-Scientist-v2 does not take any axis by a clear margin. AI-Scientist is worth picking when end-to-end experiment automation — pick this when you want a research-focused system that generates ideas, runs experiments on local GPUs, and compiles full papers automatically.
Which of AI-Scientist and AI-Scientist-v2 keeps my code private?
AI-Scientist: Self-hostable (local GPU) (4/5). AI-Scientist-v2: Cloud APIs (your keys) (3/5).
What does each one cost to run?
AI-Scientist: Requires your API keys for cloud models; local GPU needed for open-weight runs.. AI-Scientist-v2: Your API key, per-run (README cites ≈$15–$20 for experiments + ≈$5 for writing with default models).
Full profiles: SakanaAI/AI-Scientist and SakanaAI/AI-Scientist-v2. Everything else in ML experiment agents.