AI Paper Atlas

Research map / Benchmarks / NAVSIM v2

NAVSIM v2 leaderboard

11 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on NAVSIM v2. Best EPDMS: 90.6 (Ours (SFT+GRPO), NeurIPS 2026).

Numbers are as each paper reports them, read by AI from its results tables and not checked by hand. Papers can differ in setting (training data, model size, extra data); the topic column shows what each paper is about. Open the paper before you compare or quote a number.

NAVSIM v2 · EPDMS

11 papers, each with its best reported number for its own method; higher is better.

#EPDMS ↑MethodPaperTopic
190.6Ours (SFT+GRPO)
Qwen2.5-0.5B-Instruct
Grounding Driving VLA via Inverse Kinematics
NeurIPS 2026 · Table 2
Human-AI and goal-conditioned
290.1UniTeDUniTeD: Unified Temporal Diffusion for Joint Perception and Planning in Autonomous Driving
ECCV 2026 · Table 2
Recognition and long-tail
388.6HAD (C+L)HAD: Combining Hierarchical Diffusion with Metric-Decoupled RL for End-to-End Driving
ECCV 2026 · Table 2
Visual and representation RL
487.1ELF-VLA-8B
InternVL3-8B
Unleashing VLA Potentials in Autonomous Driving via Explicit Learning from Failures
CVPR 2026 · Table 2
Visual and representation RL
587.1DriveSuprim
ViT-L
DriveSuprim: Towards Precise Trajectory Selection for End-to-End Planning
AAAI 2026 · Table 3
Recognition and long-tail
687.1LaST-VLA-8B (RL)
InternVL3-8B
LaST-VLA: Thinking in Latent Spatio-Temporal Space for Vision-Language-Action in Autonomous Driving
NeurIPS 2026 · Table 2
Robot action and navigation
787.1DriveFineDriveFine: Refining-Augmented Masked Diffusion VLA for Accurate and Robust Driving
ECCV 2026 · Table 2
Robot action and navigation
886.8DriveWorld-VLADriveWorld-VLA: Unified Latent-Space World Modeling with Vision–Language–Action for Autonomous Driving
ICML 2026 · Table 2
Robot action and navigation
986.1DriveVLA-W0DriveVLA-W0: World Models Amplify Data Scaling Law in Autonomous Driving
ICLR 2026 · Table 2
Robot action and navigation
1085.8VADv2VADv2: End-to-End Autonomous Driving via Probabilistic Planning
ICLR 2026 · Table 3
Recognition and long-tail
1 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

Not affiliated with any conference, journal or the benchmark's authors. Corrections: contact@latenticlabs.com.