- Organizations
- FutureHouse
FutureHouse: AI benchmarks & research
Non-profit lab building AI scientists. Maintains LAB-Bench and BixBench.
Maintained by FutureHouse
8 benchmarks- BixBench
BixBench is a benchmark for real-world bioinformatics and computational biology data analysis. It evaluates AI models on multi-step scientific workflows that require code execution, statistical reasoning, and biological domain knowledge to interpret experimental data.
- Leader
GPT-5.580.5%
- Models scored
- 1
- BixBench MCQ
BixBench zero-shot multiple-choice accuracy on bioinformatics questions. Kept separate from the agentic BixBench workflow evaluation that requires code execution.
- Leader
Grok 4.788.4%
- Models scored
- 1
- CloningScenarios
CloningScenarios is an expert-level multi-step reasoning benchmark about difficult genetic cloning scenarios in multiple-choice format. It evaluates dual-use biological knowledge relevant to bioweapons development.
- Leader
Grok-4.1 Thinking46.0%
- Models scored
- 1
- FigQA
FigQA is a multiple-choice benchmark on interpreting scientific figures from biology papers. It evaluates dual-use biological knowledge and multimodal reasoning relevant to bioweapons development.
- Leader
Claude Mythos Preview89.0%
- Models scored
- 3
- LABBench2
LABBench2 evaluates models on real-world biology research tasks.
- Leader
Gemini 4 Argon88.8%
- Models scored
- 3
- LAB-Bench Practical
LAB-Bench practical is a public multiple-choice evaluation of everyday wet-lab and research skills, covering protocol troubleshooting, sequence reasoning, and cloning scenarios. Distinct from labbench2.
- Leader
Grok 4.776.8%
- Models scored
- 1
- ProtocolQA
ProtocolQA is a multiple-choice benchmark on troubleshooting failed experimental outcomes from common biological laboratory protocols. It evaluates dual-use biological knowledge relevant to bioweapons development.
- Leader
Grok-4.1 Thinking79.0%
- Models scored
- 1
- ProtocolQA Open-Ended
ProtocolQA Open-Ended evaluates free-form answers on troubleshooting failed experimental outcomes from common biological laboratory protocols. Kept separate from the multiple-choice ProtocolQA split.
- Leader
Grok 4.770.4%
- Models scored
- 2