EBench is a benchmark that evaluates generalist mobile manipulation policies on 26 tasks across 5 capability and 4 generalization dimensions, revealing distinct capability profiles among models with similar success rates.
hub Mixed citations
StarVLA: A Lego-like Codebase for Vision-Language-Action Model Developing
Mixed citation behavior. Most common role is background (50%).
36
Pith papers citing it
Background
50% of classified citations
abstract
Building generalist embodied agents requires integrating perception, language understanding, and action, which are core capabilities addressed by Vision-Language-Action (VLA) approaches based on multimodal foundation models, including recent advances in vision-language models and world models. Despite rapid progress, VLA methods remain fragmented across incompatible architectures, codebases, and evaluation protocols, hindering principled comparison and reproducibility. We present StarVLA, an open-source codebase for VLA research. StarVLA addresses these challenges in three aspects. First, it provides a modular backbone--action-head architecture that supports both VLM backbones (e.g., Qwen-VL) and world-model backbones (e.g., Cosmos) alongside representative action-decoding paradigms, all under a shared abstraction in which backbone and action head can each be swapped independently. Second, it provides reusable training strategies, including cross-embodiment learning and multimodal co-training, that apply consistently across supported paradigms. Third, it integrates major benchmarks, including LIBERO, SimplerEnv, RoboTwin~2.0, RoboCasa-GR1, and BEHAVIOR-1K, through a unified evaluation interface that supports both simulation and real-robot deployment. StarVLA also ships simple, fully reproducible single-benchmark training recipes that, despite minimal data engineering, already match or surpass prior methods on multiple benchmarks with both VLM and world-model backbones. To our best knowledge, StarVLA is one of the most comprehensive open-source VLA frameworks available, and we expect it to lower the barrier for reproducing existing methods and prototyping new ones. StarVLA is being actively maintained and expanded; we will update this report as the project evolves. The code and documentation are available at https://github.com/starVLA/starVLA.
hub tools
citation-role summary
background 6
baseline 2
method 2
citation-polarity summary
years
2026 36representative citing papers
DVAC uses denoising variance as an intrinsic signal to adaptively chunk actions in flow-based robot policies, improving success rates and cutting replans on LIBERO, RoboTwin, CALVIN, and real-world tasks.
The paper identifies a deployment safety gap in VLA policies where identical checkpoints can be executable-inequivalent due to action metadata mismatches, supported by a derived closed-form transform and empirical drift measurements on LIBERO benchmarks.
RotVLA models latent actions as continuous SO(n) rotations with triplet-frame supervision and flow-matching to reach 98.2% success on LIBERO and 89.6%/88.5% on RoboTwin2.0 using a 1.7B-parameter model.
LoopVLA adds recurrent refinement and learned sufficiency estimation to VLA models, cutting parameters 45% and raising throughput 1.7x while matching baseline task success on LIBERO and VLA-Arena.
GuardVLA embeds a stealthy backdoor watermark in VLAs via secret messages in visual data and uses a swap-and-detect mechanism for post-release ownership verification that preserves task performance.
Being-H0.7 adds future-aware latent reasoning to direct VLA policies via dual-branch alignment on latent queries, matching world-model benefits at VLA efficiency.
Discrete diffusion policies act as natural asynchronous executors for robotics by treating action generation as iterative unmasking, yielding higher success rates and lower computation than flow-matching real-time chunking in dynamic tasks.
ZR-0 is a dual-stream VLA model trained with dense ECoT supervision on 60M frames from 400K trajectories to enable cross-embodiment transfer in simulation and real-world settings.
T^2VLA is a test-time reinforcement learning framework for VLAs that uses internal confidence to define intrinsic rewards via similarity to high-confidence expert demonstrations and a dual-expert bootstrapping mechanism.
LA4VLA creates a 33K language-action dataset from existing demos and shows that pretraining on language-action pairs before or alongside vision-language-action training boosts success rates in sim and real robot tasks.
EventVLA introduces foundational visual anchors and a Keyframe Evidence Memory module that predicts future keyframe probabilities from VLA embeddings to improve long-horizon task success by an average of 40% on 17 simulation and 4 real-world tasks.
Qwen-RobotManip applies unified alignment across representation, motion, and behavior to enable large-scale training on heterogeneous manipulation data, yielding emergent generalization on out-of-distribution robotic benchmarks.
TempoVLA learns a single VLA policy with controllable execution speed via variable-speed trajectory augmentation and explicit speed conditioning.
ERVLA trains on a 978k-trajectory embodied CoT corpus using reasoning as supervision with dropout, then predicts actions without CoT at test time, reaching 86.9% on LIBERO-Plus and 53.2% on VLABench.
PHASER improves average success rate by up to 31% over uniform experience replay on LIBERO continual learning benchmarks for VLA models by phase-centric capacity allocation and semantic interference routing.
RoboSemanticBench reveals that representative VLA models grasp blocks successfully but select the semantically correct answer at near-random rates, indicating a gap between backbone semantics and action prediction.
FineVLA unifies robot datasets into 47k fine-grained trajectories, adds a VLM annotator and benchmark, and shows that mixing fine-grained and goal-level instructions improves steerable control without hurting task success.
Lngram is a latent N-gram conditional memory module that learns discrete symbols from hidden states for N-gram lookup, outperforming baselines in language modeling and multimodal tasks.
KeyStone improves task success rates in diffusion-based physical AI models by up to 13.3% by sampling K trajectories in parallel, clustering them in action space, and returning the medoid of the largest cluster.
LoHo-Manip enables robust long-horizon robot manipulation by using a receding-horizon VLM manager to output progress-aware subtask sequences and 2D visual traces that condition a VLA executor for automatic replanning.
A two-stage framework pretrains an action module with temporal motion priors from unconditioned trajectories using flow-matching, then transfers it to VLA training via decoder reuse and distillation, yielding better performance on cross-embodiment tasks.
An invertible adapter for flow matching enables one-step high-dimensional action generation in robotic manipulation, cutting inference time roughly in half while preserving performance.
Kairos is a native world model stack using cross-embodiment pretraining, hybrid linear temporal attention with theoretical error bounds, and deployment-aware co-design, reporting top performance on embodied benchmarks.
citing papers explorer
-
EBench: Elemental Diagnosis of Generalist Mobile Manipulation Policies
EBench is a benchmark that evaluates generalist mobile manipulation policies on 26 tasks across 5 capability and 4 generalization dimensions, revealing distinct capability profiles among models with similar success rates.
-
Denoising Tells When to Replan: Denoising-Variance Adaptive Chunking for Flow-Based Robot Policies
DVAC uses denoising variance as an intrinsic signal to adaptively chunk actions in flow-based robot policies, improving success rates and cutting replans on LIBERO, RoboTwin, CALVIN, and real-world tasks.
-
Same Weights, Different Robot: A Deployment Safety View of VLA Policies
The paper identifies a deployment safety gap in VLA policies where identical checkpoints can be executable-inequivalent due to action metadata mismatches, supported by a derived closed-form transform and empirical drift measurements on LIBERO benchmarks.
-
RotVLA: Rotational Latent Action for Vision-Language-Action Model
RotVLA models latent actions as continuous SO(n) rotations with triplet-frame supervision and flow-matching to reach 98.2% success on LIBERO and 89.6%/88.5% on RoboTwin2.0 using a 1.7B-parameter model.
-
LoopVLA: Learning Sufficiency in Recurrent Refinement for Vision-Language-Action Models
LoopVLA adds recurrent refinement and learned sufficiency estimation to VLA models, cutting parameters 45% and raising throughput 1.7x while matching baseline task success on LIBERO and VLA-Arena.
-
Towards Backdoor-Based Ownership Verification for Vision-Language-Action Models
GuardVLA embeds a stealthy backdoor watermark in VLAs via secret messages in visual data and uses a swap-and-detect mechanism for post-release ownership verification that preserves task performance.
-
Being-H0.7: A Latent World-Action Model from Egocentric Videos
Being-H0.7 adds future-aware latent reasoning to direct VLA policies via dual-branch alignment on latent queries, matching world-model benefits at VLA efficiency.
-
DiscreteRTC: Discrete Diffusion Policies are Natural Asynchronous Executors
Discrete diffusion policies act as natural asynchronous executors for robotics by treating action generation as iterative unmasking, yielding higher success rates and lower computation than flow-matching real-time chunking in dynamic tasks.
-
Training Vision-Language-Action Models with Dense Embodied Chain-of-Thought Supervision
ZR-0 is a dual-stream VLA model trained with dense ECoT supervision on 60M frames from 400K trajectories to enable cross-embodiment transfer in simulation and real-world settings.
-
Trust Your Instincts: Confidence-Driven Test-Time RL for Vision-Language-Action Models
T^2VLA is a test-time reinforcement learning framework for VLAs that uses internal confidence to define intrinsic rewards via similarity to high-confidence expert demonstrations and a dual-expert bootstrapping mechanism.
-
LA4VLA: Learning to Act without Seeing via Language-Action Pretraining
LA4VLA creates a 33K language-action dataset from existing demos and shows that pretraining on language-action pairs before or alongside vision-language-action training boosts success rates in sim and real robot tasks.
-
EventVLA: Event-Driven Visual Evidence Memory for Long-Horizon Vision-Language-Action Policies
EventVLA introduces foundational visual anchors and a Keyframe Evidence Memory module that predicts future keyframe probabilities from VLA embeddings to improve long-horizon task success by an average of 40% on 17 simulation and 4 real-world tasks.
-
Qwen-RobotManip Technical Report: Alignment Unlocks Scale for Robotic Manipulation Foundation Models
Qwen-RobotManip applies unified alignment across representation, motion, and behavior to enable large-scale training on heterogeneous manipulation data, yielding emergent generalization on out-of-distribution robotic benchmarks.
-
TempoVLA: Learning Speed-Controllable Vision-Language-Action Policies
TempoVLA learns a single VLA policy with controllable execution speed via variable-speed trajectory augmentation and explicit speed conditioning.
-
Revisiting Embodied Chain-of-Thought for Generalizable Robot Manipulation
ERVLA trains on a 978k-trajectory embodied CoT corpus using reasoning as supervision with dropout, then predicts actions without CoT at test time, reaching 86.9% on LIBERO-Plus and 53.2% on VLABench.
-
PHASER: Phase-Aware and Semantic Experience Replay for Vision-Language-Action Models
PHASER improves average success rate by up to 31% over uniform experience replay on LIBERO continual learning benchmarks for VLA models by phase-centric capacity allocation and semantic interference routing.
-
RoboSemanticBench: Diagnosing Semantic Grounding in Action Prediction for VLA Models
RoboSemanticBench reveals that representative VLA models grasp blocks successfully but select the semantically correct answer at near-random rates, indicating a gap between backbone semantics and action prediction.
-
FineVLA: Fine-Grained Instruction Alignment for Steerable Vision-Language-Action Policies
FineVLA unifies robot datasets into 47k fine-grained trajectories, adds a VLM annotator and benchmark, and shows that mixing fine-grained and goal-level instructions improves steerable control without hurting task success.
-
Lngram: N-gram Conditional Memory in Latent Space
Lngram is a latent N-gram conditional memory module that learns discrete symbols from hidden states for N-gram lookup, outperforming baselines in language modeling and multimodal tasks.
-
Geometry Guided Self-Consistency for Physical AI
KeyStone improves task success rates in diffusion-based physical AI models by up to 13.3% by sampling K trajectories in parallel, clustering them in action space, and returning the medoid of the largest cluster.
-
Long-Horizon Manipulation via Trace-Conditioned VLA Planning
LoHo-Manip enables robust long-horizon robot manipulation by using a receding-horizon VLM manager to output progress-aware subtask sequences and 2D visual traces that condition a VLA executor for automatic replanning.
-
Learning Action Priors for Cross-embodiment Robot Manipulation
A two-stage framework pretrains an action module with temporal motion priors from unconditioned trajectories using flow-matching, then transfers it to VLA training via decoder reuse and distillation, yielding better performance on cross-embodiment tasks.
-
Invertible Neural Network Adapter for One-Step Flow Matching in Robot Manipulation
An invertible adapter for flow matching enables one-step high-dimensional action generation in robotic manipulation, cutting inference time roughly in half while preserving performance.
-
Kairos: A Native World Model Stack for Physical AI
Kairos is a native world model stack using cross-embodiment pretraining, hybrid linear temporal attention with theoretical error bounds, and deployment-aware co-design, reporting top performance on embodied benchmarks.
-
World-Language-Action Model for Unified World Modeling, Language Reasoning, and Action Synthesis
WLA models use an autoregressive Transformer to jointly predict textual subtasks, subgoal images, and robot actions from instructions, images, and states, reporting SOTA success rates on RoboTwin2.0 and RMBench.
-
DeMaVLA: A Vision-Language-Action Foundation Model for Generalizable Deformable Manipulation
DeMaVLA is a VLA foundation model using a pruned action expert and flow matching, pre-trained on 5000 hours of real demonstrations and post-trained on multi-task folding data with human-in-the-loop correction, reporting competitive benchmark and real-world folding performance.
-
Rethinking VLM Representation for VLA Initialization
Experiments indicate original VLM representations are crucial for VLA performance, LoRA outperforms full finetuning, and staged robot-data pretraining yields the strongest initialization.
-
PhysBrain 1.0 Technical Report
PhysBrain 1.0 extracts scene elements, spatial dynamics, actions and depth relations from human egocentric video to create QA supervision for VLMs, then transfers the resulting physical priors to VLA policies via capability-preserving adaptation.
-
Pelican-Unify 1.0: A Unified Embodied Intelligence Model for Understanding, Reasoning, Imagination and Action
A unified embodied foundation model uses one VLM for understanding and reasoning plus a joint video-action future generator, reporting competitive scores on VLM, world modeling, and robot benchmarks without apparent compromise.
-
IntentVLA: Short-Horizon Intent Modeling for Aliased Robot Manipulation
IntentVLA conditions VLA chunk generation on a compact intent code from recent observations and introduces AliasBench to evaluate stability under short-horizon observation aliasing, reporting gains on multiple robot benchmarks.
-
Learning Action Manifold with Multi-view Latent Priors for Robotic Manipulation
The method uses multi-view diffusion priors and action manifold learning to resolve depth ambiguity and improve action prediction in VLA robotic manipulation models, reporting higher success rates than baselines on LIBERO, RoboTwin, and real-robot tasks.
-
VLA-GSE: Boosting Parameter-Efficient Fine-Tuning in VLA with Generalized and Specialized Experts
VLA-GSE uses spectral decomposition of the VLA backbone to create generalized and specialized experts, enabling effective robot task adaptation while updating only 2.51% of parameters and achieving 81.2% zero-shot success on LIBERO-Plus.
-
VLA Foundry: A Unified Framework for Training Vision-Language-Action Models
VLA Foundry provides a single training stack for VLA models and releases open models that match prior closed-source performance or outperform baselines on multi-task manipulation in simulation.
-
RLDX-1 Technical Report
RLDX-1 outperforms frontier VLAs such as π0.5 and GR00T N1.6 on dexterous manipulation benchmarks, reaching 86.8% success on ALLEX humanoid tasks versus around 40% for the baselines.
-
Agentic World Modeling: Foundations, Capabilities, Laws, and Beyond
A survey proposing a three-level capability taxonomy (L1 Predictor, L2 Simulator, L3 Evolver) for world models across physical, digital, social, and scientific domains.
-
JoyAI-RA 0.1: A Foundation Model for Robotic Autonomy
JoyAI-RA is a multi-source pretrained VLA model that claims to bridge human-to-robot embodiment gaps via data unification and outperforms prior methods on generalization-heavy robotic tasks.