CoLT-Drive - Road Hazard: leaderboard

The road hazard category of CoLT-Drive (Tang et al., 2026): road-surface hazards such as potholes, open manholes, oil spills and cables inserted into front-view driving scenes with the surrounding traffic kept, where the expected response adapts to the hazard. The model outputs a longitudinal and lateral meta-action, and the score is pair accuracy in percent against the human-adjudicated acceptable set; higher is better.

Metric: Pair accuracy (%; the 348 v_full samples, naturalistic surrounding traffic kept, whose inserted object is one of the road-surface hazards (potholes, open manholes, oil spills, glass and cables), where the model outputs an acceptable longitudinal-lateral meta-action pair for a front-view driving image with a counterfactually inserted rare object, ego-motion history and navigation command; human-adjudicated acceptable-pair sets; same structured prompt and greedy decoding for every model; decisions extracted by rule and normalized by a text-only DeepSeek-V4-Pro mapper, then scored deterministically; invalid outputs count as wrong). Source: arxiv.org. Saturation forecast: Around February 2027. 9 models tracked.

Top models

No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.

Interactive version: theaggregate.ai/benchmark?slug=colt-drive-road-hazard · How It Works · Data refreshed daily, snapshot 2026-11-12.