Modelpublic

InternRobotics/InternVLA-A1-3B-RoboTwin

sourceHugging Facecc-by-nc-sa-4.0updated 7mo agoView on Hugging Face

15likes27downloads

Model Card

InternVLA-A1: Unifying Understanding, Generation and Action for Robotic Manipulation

![Paper](https://arxiv.org/pdf/2601.02456) ![Code](https://github.com/InternRobotics/InternVLA-A1) ![Data](https://huggingface.co/datasets/InternRobotics/InternData-A1) ![Website](https://internrobotics.github.io/internvla-a1.github.io/)

<strong>InternVLA-A1</strong> integrates understanding, generation, and action experts via a Mixture-of-Transformers (MoT) framework, which synergizes MLLMs' semantic reasoning with world-model-style dynamics prediction to guide action execution.

Building upon InternVL3 and Qwen3-VL, we instantiate InternVLA-A1 at 2B and 3B parameter scales. Covering different model scales and pre-training data configurations, we release the InternVLA-A1 series:

—[x] InternVLA-A1-3B: pretrained on heterogeneous data sources over real-robot data, synthetic simulation data, and human videos
—[x] InternVLA-A1-3B-RoboTwin: finetuned on RoboTwin 2.0 benchmark
—[ ] InternVLA-A1-3B-Pretrain-InternData-A1: pretrained on InternData-A1 only
—[ ] InternVLA-A1-2B-Pretrain-InternData-A1: pretrained on InternData-A1 only

Evaluation on RoboTwin 2.0 Simulation Benchmark

Setting: All models are jointly fine-tuned across 50 tasks (50 clean + 500 randomized demos each).

Performance Summary: InternVLA-A1-3B achieves the highest success rates across both Easy and Hard settings on the RoboTwin 2.0 Benchmark (averaged over 50 tasks).

Metric	pi0	pi0.5	InternVLA-A1-3B
Avg. Success (Easy)	79.98%	86.76%	89.40%
Avg. Success (Hard)	79.50%	86.96%	89.64%

🔑 Key Features

—🔮 The Core: Synergizes MLLM's semantic understanding with world-model-style dynamic prediction, enabling it to "imagine" the future and guide adaptive actions.
—🚀 The Fuel: Enables joint training on heterogeneous data sources over real-world robot data, synthetic simulation data, and egocentric human videos.
—⚡ The Output: Tackles highly dynamic scenarios with effortless mastery.

Usage

Please refer to our official repo InternVLA-A1.

Demonstrations

InternVLA-A1 exhibits consistent robustness across static manipulation, dynamic manipulation, and simulation benchmarks, especially demonstrating remarkable superiority in dynamic scenarios.

⚡ Dynamic Manipulation Tasks

<div style="display: flex; flex-direction: column; align-items: center; gap: 5px;">  <div style="display: flex; justify-content: center; align-items: center; gap: 5px;"> <video controls autoplay loop muted width="200" style="border-radius: 5px; box-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);"> <source src="https://huggingface.co/spaces/Jia-Zeng/InternVLAA1Media/resolve/main/expresssortingcomplete.mp4" type="video/mp4"> </video> <video controls autoplay loop muted width="200" style="border-radius: 5px; box-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);"> <source src="https://huggingface.co/spaces/Jia-Zeng/InternVLAA1Media/resolve/main/In-motionIngredientPicking4x.mp4" type="video/mp4"> </video> <video controls autoplay loop muted width="200" style="border-radius: 5px; box-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);"> <source src="https://huggingface.co/spaces/Jia-Zeng/InternVLAA1Media/resolve/main/expresssortingupright3.mp4" type="video/mp4"> </video> </div>  <div style="display: flex; justify-content: center; align-items: center; gap: 5px;"> <video controls autoplay loop muted width="200" style="border-radius: 5px; box-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);"> <source src="https://huggingface.co/spaces/Jia-Zeng/InternVLAA1Media/resolve/main/expresssortinginverted2.mp4" type="video/mp4"> </video> <video controls autoplay loop muted width="200" style="border-radius: 5px; box-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);"> <source src="https://huggingface.co/spaces/Jia-Zeng/InternVLAA1Media/resolve/main/expresssortingupright2.mp4" type="video/mp4"> </video> <video controls autoplay loop muted width="200" style="border-radius: 5px; box-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);"> <source src="https://huggingface.co/spaces/Jia-Zeng/InternVLAA1Media/resolve/main/expresssortingupright_1.mp4" type="video/mp4"> </video> </div> <p><em>InternVLA-A1 exhibits exceptional robustness in highly dynamic scenarios.</em></p> </div>

🤖 Static Manipulation Tasks

<div style="display: flex; flex-direction: column; align-items: center; gap: 5px;">  <div style="display: flex; justify-content: center; align-items: center; gap: 5px;"> <video controls autoplay loop muted width="200" style="border-radius: 5px; box-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);"> <source src="https://huggingface.co/spaces/Jia-Zeng/InternVLAA1Media/resolve/main/zigbag4x.mp4" type="video/mp4"> </video> <video controls autoplay loop muted width="200" style="border-radius: 5px; box-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);"> <source src="https://huggingface.co/spaces/Jia-Zeng/InternVLAA1Media/resolve/main/sortparts4x.mp4" type="video/mp4"> </video> <video controls autoplay loop muted width="200" style="border-radius: 5px; box-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);"> <source src="https://huggingface.co/spaces/Jia-Zeng/InternVLAA1Media/resolve/main/unscrewcap4x.mp4" type="video/mp4"> </video> </div>  <div style="display: flex; justify-content: center; align-items: center; gap: 5px;"> <video controls autoplay loop muted width="200" style="border-radius: 5px; box-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);"> <source src="https://huggingface.co/spaces/Jia-Zeng/InternVLAA1Media/resolve/main/wipestain4x.mp4" type="video/mp4"> </video> <video controls autoplay loop muted width="200" style="border-radius: 5px; box-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);"> <source src="https://huggingface.co/spaces/Jia-Zeng/InternVLAA1Media/resolve/main/placeflower4x.mp4" type="video/mp4"> </video> <video controls autoplay loop muted width="200" style="border-radius: 5px; box-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);"> <source src="https://huggingface.co/spaces/Jia-Zeng/InternVLAA1Media/resolve/main/sweeptrash4x.mp4" type="video/mp4"> </video> </div> <p><em>InternVLA-A1 demonstrates superior proficiency in dexterous and fine-grained manipulation.</em></p> </div>

</div>

License and Citation

All the code within this repo are under CC BY-NC-SA 4.0. Please consider citing our project if it helps your research.

BibTeX

@article{internvla_a1,
  title={InternVLA-A1: Unifying Understanding, Generation and Action for Robotic Manipulation},
  author={Cai, Junhao and Cai, Zetao and Cao, Jiafei and Chen, Yilun and He, Zeyu and Jiang, Lei and Li, Hang and Li, Hengjie and Li, Yang and Liu, Yufei and others},
  journal={arXiv preprint arXiv:2601.02456},
  year={2026}
}