AI Paper Atlas

Research map / Benchmarks / Bench2Drive closed-loop

Bench2Drive closed-loop leaderboard

24 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on Bench2Drive closed-loop. Best Driving Score: 89.42 (AutoMoT + (with action refiner), ICML 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.

Bench2Drive closed-loop · Driving Score

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

#Driving Score ↑MethodPaperTopic
189.42AutoMoT + (with action refiner)
Qwen3-VL-4B + 1.6B action expert
AutoMoT: A Unified Vision-Language-Action Model with Asynchronous Mixture -of-Transformers for End-to-End Autonomous Driving
ICML 2026 · Table 1
Robot action and navigation
288.42KnowVal-SimLingo
SimLingo
KnowVal: A Knowledge-Augmented and Value-Guided Autonomous Driving System
CVPR 2026 · Table 2
Agents, code and math
387.98LADA (5% HLC, 5% CF)
SimLingo (InternVL2-1B)
Less Language, More Latents: Annotation-Efficient VLAs for Driving
NeurIPS 2026 · Table 1
Robot action and navigation
486.28R2SEReinforced Refinement With Self-Aware Expansion for End-to-End Autonomous Driving
TPAMI 2026 · Table 2
Visual and representation RL
584.87SimLingo+SA
SimLingo (InternVL2-1B)
Scene-Adaptive VLA: Efficient Autonomous Driving via Dynamic Layer Routing
NeurIPS 2026 · Table 2
Robot action and navigation
681.4OmniSpace
Qwen3-VL 4B
OmniSpace: Efficient Geometry Awareness for Autonomous Vehicles MLLMs
NeurIPS 2026 · Table 2
Multimodal LLMs
780.59MindDrive-L (online RL)
Qwen2.5-3B
MindDrive: A Vision-Language-Action Model for Autonomous Driving via Online Reinforcement Learning
ECCV 2026 · Table 1
Robot action and navigation
879.23WPT-TeacherWPT: World-to-Policy Transfer via Online World Model Distillation
CVPR 2026 · Table 2
Reasoning and chain of thought
979.14PixelPilot
Qwen2.5-VL-7B
PixelPilot: Scalable Vision-Language-Action Models for End-to-End Autonomous Driving
ECCV 2026 · Table 4
Robot action and navigation
1078.21CogDriver-AgentCogDriver: Integrating Cognitive Inertia for Temporally Coherent Planning in Autonomous Driving
CVPR 2026 · Table 2
Robot action and navigation
13 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

Bench2Drive closed-loop · success rate

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

#success rate ↑MethodPaperTopic
174.09AutoMoT + (with action refiner)
Qwen3-VL-4B + 1.6B action expert
AutoMoT: A Unified Vision-Language-Action Model with Asynchronous Mixture -of-Transformers for End-to-End Autonomous Driving
ICML 2026 · Table 1
Robot action and navigation
270.46LADA (5% HLC, 5% CF)
SimLingo (InternVL2-1B)
Less Language, More Latents: Annotation-Efficient VLAs for Driving
NeurIPS 2026 · Table 1
Robot action and navigation
369.03KnowVal-SimLingoKnowVal: A Knowledge-Augmented and Value-Guided Autonomous Driving System
CVPR 2026 · Table 2
Agents, code and math
465.52OmniSpace
Qwen3-VL 4B
OmniSpace: Efficient Geometry Awareness for Autonomous Vehicles MLLMs
NeurIPS 2026 · Table 2
Multimodal LLMs
565.45SimLingo+SA
SimLingo (InternVL2-1B)
Scene-Adaptive VLA: Efficient Autonomous Driving via Dynamic Layer Routing
NeurIPS 2026 · Table 2
Robot action and navigation
658.87PixelPilot
Qwen2.5-VL-7B
PixelPilot: Scalable Vision-Language-Action Models for End-to-End Autonomous Driving
ECCV 2026 · Table 4
Robot action and navigation
758.34Expert-VADv2ExpertAD: Enhancing Autonomous Driving Systems with Mixture of Experts
AAAI 2026 · Table 1
Recognition and long-tail
858.26MindDrive-L (online RL)
Qwen2.5-3B
MindDrive: A Vision-Language-Action Model for Autonomous Driving via Online Reinforcement Learning
ECCV 2026 · Table 1
Robot action and navigation
956.93CogDriver-AgentCogDriver: Integrating Cognitive Inertia for Temporally Coherent Planning in Autonomous Driving
CVPR 2026 · Table 2
Robot action and navigation
1055.11SpaceDrive+ (ours)SpaceDrive: Infusing Spatial Awareness into VLM-based Autonomous Driving
CVPR 2026 · Table 2
Multimodal LLMs
12 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.