Benchmark RadarAI BENCHMARK PROFILE
ABC-Bench
ABC-Bench evaluates LLM agents on biosecurity-relevant tasks including liquid handling robot code generation, DNA fragment design, and DNA synthesis screening evasion, with wet-lab validation.
- Released
- 2026-06-09
- Readiness
- Paper only
- Primary field
- Health & Life Sciences
Why it matters
Measures agentic AI capabilities relevant to biosecurity, offering a standardized protocol to assess dual-use risks and inform safeguards in biological research contexts.
Motivation
Large language models (LLMs) are rapidly acquiring capabilities relevant to biological research, from literature synthesis to interpretation of experimental data.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.