# Matbench Discovery

Matbench Discovery is an AI benchmark; this page records what it measures and traces its data, grader, and scores to primary sources.

Compares the accuracy, robustness, and computational cost of models used for atomistic and materials-discovery tasks.

## Details

- Address: Online, Global
- Website: [matbench-discovery.materialsproject.org/](<https://matbench-discovery.materialsproject.org/>)
- Tags: Helpful, Science
- Category: Science
- Status: Active
- Data licence: MIT
- Data access: open
- Grader: exact
- Organization: Matbench Discovery
- Released: 2022-06
- Paper: [www.nature.com/articles/s42256-025-01055-1](<https://www.nature.com/articles/s42256-025-01055-1>)
- Leaderboard: [matbench-discovery.materialsproject.org/](<https://matbench-discovery.materialsproject.org/>)
- Repository: [github.com/janosh/matbench-discovery](<https://github.com/janosh/matbench-discovery>)
- Top-score provenance: unknown
- ABI family: No

## Questions

### Where is Matbench Discovery?

Matbench Discovery is at Online, Global.

### How do I contact Matbench Discovery?

To contact Matbench Discovery, visit https://matbench-discovery.materialsproject.org/.

## Source

From [Benchmark primary sources](https://aibenchmarkindex.com/about), license CC0-1.0, checked 2026-10-15.

Last updated October 15, 2026.

## Links

- This benchmark on the [web page](<https://aibenchmarkindex.com/global/online/matbench-discovery>)
- [Science in Online](<https://aibenchmarkindex.com/global/online/category/science.md>)
- [More benchmarks in Online](<https://aibenchmarkindex.com/global/online.md>)

> **Check before you compare.** Benchmark methods and leaderboards change. Follow the cited primary source before using a result in a decision.