# SafetyBench

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

Evaluates large-language-model safety knowledge with Chinese and English multiple-choice questions across seven risk categories.

## Details

- Address: Online, Global
- Website: [llmbench.ai/safety](<https://llmbench.ai/safety>)
- Tags: Helpful, Science
- Category: Safety
- Status: Active
- Data licence: MIT
- Data access: open
- Grader: exact
- Organization: Tsinghua University CoAI
- Released: 2023-09
- Paper: [arxiv.org/abs/2309.07045](<https://arxiv.org/abs/2309.07045>)
- Leaderboard: [llmbench.ai/safety](<https://llmbench.ai/safety>)
- Repository: [github.com/thu-coai/SafetyBench](<https://github.com/thu-coai/SafetyBench>)
- Task count: 11,435
- Top-score provenance: unknown
- ABI family: No

## Questions

### Where is SafetyBench?

SafetyBench is at Online, Global.

### How do I contact SafetyBench?

To contact SafetyBench, visit https://llmbench.ai/safety.

## 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/safetybench>)
- [Safety in Online](<https://aibenchmarkindex.com/global/online/category/safety.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.