Bench2Drive Leaderboard

June 26, 2026 · View on GitHub

Bench2Drive is a popular autonomous driving benchmark. Over 60 methods have been tested on Bench2Drive. Papers using the benchmark usually only report a few baselines due to space constraints making it hard to keep an overview over the literature. To solve this, this repository contains a complete list of all methods that have currently been tested on Bench2Drive.

The results are from Bench2Drive version 0.0.3 and self-reported, copied from the respective publications. Methods are sorted by Driving Score (DS). At the bottom of the table, separated by a line, are privileged methods that use ground truth perception as input. This leaderboard compares the final performance of entire methods, including data acquisition methods, not just architectures. Methods vary among many axes: driving dataset used, pre-training strategy, training compute, inference compute, number of parameters, IL/RL, etc. In particular, many (6/8) of the current top-performing methods use PDM-Lite or LEAD as expert.

To add a new method please open a pull request. For fixing errors please open a GitHub issue.

Main.tex contains the leaderboard table as latex with the associated citations in CVPR-style.

Last updated: 24. Jun. 2026

Results:

MethodDS ↑SR ↑YearVenue
AD-MLP18.050.002023-
TCP40.7015.002022NeurIPS
ReAL-AD41.1711.362025ICCV
VAD42.3515.002023ICCV
SparseDrive44.5416.712025ICRA
GenAD44.8115.902024ECCV
Dual-AEB45.2310.002025ICRA
FUMP45.6716.362025-
FocalAD45.7717.302025-
UniAD45.8116.362023CVPR
MomAD47.9118.112025CVPR
CogAD48.3024.002025-
AdaptiveAD49.4719.232025-
CausalVAD49.8319.422026CVPR
BridgeAD50.0622.732025CVPR
E³AD50.0720.122025NeurIPS
X-Driver51.7018.102025-
DiFSD52.0221.002024-
SpaRC-AD55.6030.002025Workshop
SeerDrive58.3230.172025NeurIPS
WoTE61.7131.362025ICCV
DriveDPO62.0230.622025NeurIPS
ThinkTwice62.4431.232022CVPR
DriveTransformer-L.63.4635.012025ICLR
ReasonPlan64.0134.552025CoRL
DriveAdapter64.2233.082023ICCV
VeteranAD64.2233.852025-
Drive-JEPA64.5236.822026-
iPad65.0235.912025-
StuckSolver65.2336.322025-
GEMINUS65.3937.732025-
HiPro-AD65.4837.312025Sensors
MindDriver65.4839.552026-
VDRive66.2550.512025-
RAP-ResNet66.4237.272026ICLR
CAPS66.7652.872025-
SafeDrive66.7742.402026CVPR
DriveMamba66.8237.732026ICLR
SNG67.1735.902025-
Diffusion-Phy67.4936.512026PR
DiffAD67.9238.642025-
OAIAD68.7348.862025Machines
DIVER68.9036.752025-
ReCogDrive71.3645.452025-
Raw2Drive71.3650.242025NeurIPS
TakeAD71.3940.832025RA-L
VL73.2965.442025ICRA
Hydra-NeXt73.8650.002025ICCV
DriveMoE74.2248.642025-
ETA74.3348.332025ICCV
GuideFlow75.2151.362025-
DiffusionDrive-temp77.6852.722025CVPR
ORION77.7454.622025ICCV
SpaceDrive+78.0255.112025CVPR
MindDrive78.0455.092025CVPR
PGS78.0848.642025NeurIPS
Expert-VADv278.1858.342026AAAI
PerlAD78.7057.272026RA-L
UniDriveVLA78.7351.822026-
AutoVLA78.8457.732025NeurIPS
VLGA79.0852.732026-
GaussianFusion79.1054.402025NeurIPS
UniDrive-WM79.2256.362026-
Cot4AD-CoT81.2255.782025-
GeRo81.9060.102026-
DMW82.7271.562026CVPR
TFM+void83.6964.852025NIK
TF++84.2167.272024-
SimLingo85.0767.272025CVPR
R2SE86.2869.542026PAMI
DeLL86.6668.632026-
HiP-AD86.7769.092025ICCV
BridgeDrive86.8772.272026ICLR
DiffRefiner87.171.42025-
CaAD87.5371.812026-
CriticVLA88.0273.332026-
BevAD88.1172.732026CVPRF
MAPLE88.370.52026-
RaWMPC88.3170.482026-
DynVLA88.3472.732026-
KnowVal88.4269.032025-
GraphPilot88.8972.732025-
VECTOR-DRIVE88.9171.822026-
PersonaDrive88.9572.292026-
AlignDrive89.0773.182026-
VLADriver-RAG89.1270.422026-
SparseDriveV289.1570.002026-
AutoMoT89.4274.092026ICML
TakeVLA89.7273.732026-
BLUE90.5876.182026-
SteerVLA91-2026-
LinkVLA91.0174.552026-
TFv695.2886.812026CVPR
---------------
Think2Drive91.8585.412024ECCV
LEAD96.7896.592026CVPR
PDM-Lite97.0292.272024ECCV