BioStudyBench: Evaluating Agents on Post-Cutoff Biomedical Studies
cs.AI
Submitted: 2026-10-06
Updated: 2026-10-06
Comments: Accepted into AgenticLS (NeurIPS 2026 workshop)
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Sources
- ScienceAgentBench: Toward Rigorous Assessment of Language Agents for Data-Driven Scientific Discovery
- Can Prompts Rewind Time for LLMs? Evaluating the Effectiveness of Prompted Knowledge Cutoffs
- BLADE: Benchmarking Language Model Agents for Data-Driven Science
- Can AI Agents Synthesize Scientific Conclusions?
- BixBench3: Benchmarking AI agents on research-study-scale computational biology tasks
- Can LLM Agents Generate Real-World Evidence? Evaluating Observational Studies in Medical Databases
- The Memorization Problem: Can We Trust LLMs' Economic Forecasts?
- DiscoveryBench: Towards Data-Driven Discovery with Large Language Models
- BixBench: a Comprehensive Benchmark for LLM-based Agents in Computational Biology
- CORE-Bench: Fostering the Credibility of Published Research Through a Computational Reproducibility Agent Benchmark
- PaperBench: Evaluating AI's Ability to Replicate AI Research
- FIRE-Bench: Evaluating AI Agents on the Rediscovery of Scientific Insights
- BioDSA-1K: Benchmarking Data Science Agents for Biomedical Research
- LiveBench: A Challenging, Contamination-Limited LLM Benchmark
- AntiLeakBench: Preventing Data Contamination by Automatically Constructing Benchmarks with Updated Real-World Knowledge
- Benchmarking Benchmark Leakage in Large Language Models
- Proof of Time: A Benchmark for Evaluating Scientific Idea Judgments
- ReplicationBench: Can AI Agents Replicate Astrophysics Research Papers?
Related papers
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Agents' Last Exam
- The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
- MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
- Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems