5arXiv cs.CL (Computation and Language)·
MedFailBench: Clinician-built benchmark for medical AI safety boundary failures
MedFailBench is a new open-source benchmark designed to identify which safety boundary a medical AI system failed, rather than measuring correctness. The v0.2.1 release contains 44 clinician-reviewed synthetic cases annotated by severity (1–5) and six safety gate types including missed urgent escalation, unsafe dosing, and evidence fabrication. The benchmark includes a HuggingFace leaderboard preview, a safety gate taxonomy, and an automated pipeline for archiving model-response screening runs, released under Apache-2.0 and CC-BY-4.0.
Related guides (3)
Related events (8)
Introducing HealthBench
OpenAI has released HealthBench, a new evaluation benchmark designed to assess AI model performance and safety in healthcare settings. The benchmark was developed with input from over 250 physicians and targets realistic clinical scenarios. It aims to establish a shared standard for measuring how well AI models handle health-related tasks.
MedRealMM: Real-world multimodal benchmark for Chinese online medical consultation
Researchers introduce MedRealMM, a benchmark of 5,620 real-world multimodal patient-doctor interaction cases drawn from a nationwide Chinese internet hospital, spanning 64 clinical departments. The benchmark uses a Multimodal Clinical Challenge Point (MCCP) framework to identify clinically demanding moments and evaluates 19 general-purpose and medical-specialized LLMs. Key findings show that frontier models still fall below online physician performance, particularly on safety-sensitive error avoidance, and that image information is critical for reliable clinical performance. The dataset will be publicly released on Hugging Face.
OpenAI introduces LifeSciBench, a life sciences AI evaluation benchmark
OpenAI has released LifeSciBench, a benchmark designed to evaluate AI systems on real-world life science research tasks and decisions. The benchmark is described as expert-authored and expert-reviewed, targeting domain-specific evaluation in biology and related fields. This addresses a gap in specialized scientific benchmarking for AI systems.
MedMisBench: LLMs show fragile epistemic resilience under misleading medical context
Researchers introduce MedMisBench, a benchmark of 10,932 medical questions paired with 48,889 misleading context injections, to measure whether LLMs maintain correct medical judgment under adversarial pressure. Across 11 model configurations, mean accuracy drops from 71.1% to 38.0% when misleading context is injected, with authority-framed falsehoods achieving 69.5% attack success. A 14-member international clinical panel flagged serious potential harm in 38.2% of reviewed cases. The work argues that existing medical benchmarks measure knowledge but not robustness to manipulation, exposing a structural gap in LLM safety evaluation for healthcare.
EMPATH: Multilingual multi-turn safety benchmark for emotional-support chatbots reveals score inflation and run-to-run reliability failures
EMPATH is a new arXiv benchmark for evaluating the safety of emotional-support chatbots, using an auditor model to generate multi-turn crisis conversations and a calibrated judge model to score transcripts across 19 metrics in five dimensions. Built for Mexican Spanish and US English, the benchmark surfaces score inflation on 10 of 19 metrics under uncalibrated rubrics and finds that run-to-run reliability is a per-model safety property: one model swings 2–10 points on a crisis metric across identical reruns, and DeepSeek V4 Pro produces different conversations at temperature 0. Evaluation of three frontier models shows aggregate scores within 0.74 points but per-metric divergences up to six points, with rankings stable across a cross-family judge at 93% within ±1.
ClinEnv: Interactive Multi-Stage Long-Horizon EHR Benchmark for Clinical Agent Evaluation
ClinEnv is a new interactive benchmark that evaluates LLMs as attending physicians over real inpatient admissions using a Longitudinal Inpatient Simulation paradigm. Each case is decomposed into sequential decision stages where models must query four specialized agents before committing to medications, procedures, and diagnoses. Across seven evaluated models, the best achieves only 0.31 decision F1, with a sharp gap between diagnosis recovery (0.51 F1) and management actions (0.17 F1). The benchmark uniquely measures information-acquisition process quality alongside outcome quality, exposing a gap invisible to static or outcome-only evaluations.
OmniaBench: A broad diagnostic benchmark for general AI agents across 354 domains
OmniaBench is a new benchmark for evaluating general AI agents across diverse scenarios, spanning 90 level-1 and 354 level-2 domains derived from app stores, product documents, and web retrieval. The benchmark contains 1,431 tasks with single-turn and multi-turn formats, a ten-dimensional capability taxonomy, and eight atomic difficulty factors for fine-grained analysis. Even frontier models like Claude Sonnet 5 and GPT-5.6-Sol achieve only ~58% Overall Pass@1, revealing persistent weaknesses in planning, constraint maintenance, and adaptive correction. The work addresses a gap in existing agent benchmarks that tend to cover narrow tool ecosystems or interaction formats.
MLE-bench: Evaluating Machine Learning Agents on Machine Learning Engineering
OpenAI introduces MLE-bench, a benchmark designed to measure AI agent performance on machine learning engineering tasks. The benchmark draws from Kaggle competitions to evaluate agents on realistic ML engineering workflows. Initial results show that current agents, including those powered by o1-preview, achieve competitive performance on a subset of tasks but fall well short of top human competitors. The benchmark is intended to track progress in agentic ML capabilities over time.