Open Call for Contributions

Submit Research Tasks,
Become a Co-Author

Help us build the most comprehensive benchmark for AI-driven scientific research. Contribute quality tasks and earn recognition as a co-author on the ResearchClawBench paper.

Submit Your Tasks Now Learn More

Contribution Rewards

Your contributions directly impact the future of AI research evaluation

1 Task

Acknowledgment in the paper

  • Listed in acknowledgments
  • GitHub contributor badge

3+ Tasks

Co-Author Status

  • Paper co-authorship
  • Full attribution
  • Research collaboration

10+ Tasks

Core Contributor

  • All co-author benefits
  • Priority collaboration
  • Future project involvement

How It Works

Simple 4-step process to contribute and earn recognition

1

Design Your Research Task

Create a research task from your domain expertise. Tasks should be specific, well-defined scientific problems that can be solved through a research workflow (literature review, hypothesis, experiment, analysis).

2

Submit via Hugging Face

Use our Hugging Face Spaces submission portal to upload your task. Include task description, expected methodology, evaluation criteria, and any relevant datasets or resources.

3

Quality Review

Our team reviews submissions for scientific rigor, clarity, and suitability for AI agent evaluation. We may reach out for clarifications or improvements.

4

Earn Recognition

Once accepted, your task becomes part of ResearchClawBench. Reach the contribution threshold to become a co-author on our benchmark paper!

What Tasks Do We Need?

We welcome research tasks across all scientific domains

Domains of Interest

Life Sciences Physics Chemistry Materials Science Earth Sciences Astronomy AI/ML Research Computer Science Mathematics Biomedical Engineering Economics

Good Task Examples

  • Reproduce a specific paper's experimental results
  • Propose improvements to an existing algorithm
  • Analyze a dataset and derive scientific insights
  • Design and validate a new hypothesis
  • Survey and synthesize literature on a topic

What We Don't Accept

  • Vague or overly broad research questions
  • Tasks requiring proprietary/inaccessible data
  • Simple Q&A or trivia questions
  • Tasks with no verifiable ground truth
  • Ethically problematic research directions

Frequently Asked Questions

How long does the review process take?

Typically 1-2 weeks. We'll notify you via the email provided during submission.

Can I submit tasks from my ongoing research?

Yes! Tasks derived from your research are welcome, as long as they're suitable for AI agent evaluation and don't conflict with any publication embargoes.

What if my task needs revision?

We'll provide feedback and work with you to refine the task. Revised submissions still count toward your contribution total.

When will the paper be published?

We're targeting submission to a major venue in 2026. Co-authors will be notified of publication timeline and involved in the review process.

Can teams submit together?

Yes! Team submissions are welcome. Each accepted task counts toward all listed contributors.

Submissions Open Now

Ready to Contribute?

Join researchers worldwide in building the definitive benchmark for AI scientific research capabilities.

Submit Your Tasks

Questions? Contact us at researchclawbench@internscience.org