DrivingGen: A Comprehensive Benchmark for Generative Video World Models in Autonomous Driving

TL;DR

DrivingGen introduces a comprehensive benchmark for evaluating generative video world models in autonomous driving, addressing current limitations in metrics, datasets, and evaluation scope to advance safe and realistic simulation.

Contribution

It provides the first diverse dataset and suite of metrics for assessing generative driving world models, enabling more reliable and controllable autonomous driving simulations.

Findings

General models excel in visual quality but lack physical accuracy.

Driving-specific models better capture motion but have lower visual realism.

Benchmark reveals trade-offs between visual fidelity and physical plausibility.

Abstract

Video generation models, as one form of world models, have emerged as one of the most exciting frontiers in AI, promising agents the ability to imagine the future by modeling the temporal evolution of complex scenes. In autonomous driving, this vision gives rise to driving world models: generative simulators that imagine ego and agent futures, enabling scalable simulation, safe testing of corner cases, and rich synthetic data generation. Yet, despite fast-growing research activity, the field lacks a rigorous benchmark to measure progress and guide priorities. Existing evaluations remain limited: generic video metrics overlook safety-critical imaging factors; trajectory plausibility is rarely quantified; temporal and agent-level consistency is neglected; and controllability with respect to ego conditioning is ignored. Moreover, current datasets fail to cover the diversity of conditions required for real-world deployment. To address these gaps, we present DrivingGen, the first comprehensive benchmark for generative driving world models. DrivingGen combines a diverse evaluation dataset curated from both driving datasets and internet-scale video sources, spanning varied weather, time of day, geographic regions, and complex maneuvers, with a suite of new metrics that jointly assess visual realism, trajectory plausibility, temporal coherence, and controllability. Benchmarking 14 state-of-the-art models reveals clear trade-offs: general models look better but break physics, while driving-specific ones capture motion realistically but lag in visual quality. DrivingGen offers a unified evaluation framework to foster reliable, controllable, and deployable driving world models, enabling scalable simulation, planning, and data-driven decision-making.