LabRobFail-Bench - Failure Detection: leaderboard

Metric: Top-1 accuracy (%; binary detection of whether the trajectory contains a failure; robot-manipulation videos from simulated chemical self-driving laboratories with injected control, physics and semantic failures, seen-environment test split; answers from simulator ground truth; zero-shot VLM). Source: arxiv.org. Saturation forecast: Around 2028. 7 models tracked.

Top models

#ModelScore
1Gemini 2.5 Flash57.13
2Gemini 2.0 Flash56.85
3GPT-5.452.34

Interactive version: theaggregate.ai/benchmark?slug=labrobfail-bench-failure-detection · How It Works · Data refreshed daily, snapshot 2026-09-29.