Request access to Verne V0.5

Verne reviews each request before it gives access. Verne uses your answers to manage your license and to contact you about Verne V0.5. The Verne Privacy Notice describes how Verne processes this information.

The Verne V0.5 Model License, version 1.0, governs the use of Verne V0.5. Read the license before you request access. When you submit this form, you accept the license. For commercial use, please contact partnership@vernerobotics.com. We would love to work with you! The license does not permit use in a country or region under comprehensive US sanctions, or use by a person on a US restricted-party list.

Verne V0.5 base

Verne V0.5 is an open-weight world action model for general-purpose robotics, made by Verne. It learns to predict both future observations and robot actions. Learning how a scene changes helps the model learn how to act in it.

This is the pretrained base checkpoint. It is the start point for fine-tuning on demonstrations from your robot and task.

Model details


Developed by	Verne Team
Model type	World action model (WAM): a video transformer and an action transformer, trained with flow matching
Inputs	Current camera frames, memory of past observations, robot state, and a language instruction
Output	A chunk of 32 targets at 30 Hz: tool poses, gripper openings, base poses, and progress
Instruction language	English
Parameters	8.2 billion in `model.safetensors`. The Wan2.2 VAE and text encoder load separately.
Initialized from	Wan2.2-TI2V-5B, with a deeper video transformer and a new action transformer, then pretrained by Verne
License	Verne V0.5 Model License

Checkpoints

Checkpoint	Description
v0.5-base	Pretrained base. Start here to fine-tune.
v0.5-libero	Fine-tuned on LIBERO. 99.3% average success.
v0.5-robotwin	Fine-tuned on RoboTwin 2.0. 81.0% clean, 73.4% randomized.

Model

A video transformer and an action transformer share information through masked attention. During training, they learn together with video and action flow-matching objectives. The video objective teaches the model how scenes change. The action objective teaches it the motion that the robot must do.

At runtime, only the action path runs. The model does not generate future video. It reads the current camera frames, its memory of past observations, the robot state, and a language instruction. It outputs a chunk of actions at 30 Hz.

Training data

We pretrained V0.5 on over 1.7 million demonstrations from four sources:

Third-party vendors.
Open-source datasets.
Verne teleoperation, in the robot's own embodiment and action space.
Verne handheld grippers.

Before training, we check the recording timing and the motion quality, cut bad segments, and align each language instruction with the demonstrated behavior.

Training procedure

V0.5 trains in three stages. This checkpoint completes stages 1 and 2.

Robot-video pretraining. Adapt a video model to robot observations, so it learns how manipulation changes a scene.
Joint video and action pretraining. Train the video and action transformers together on paired observations and actions.
Task post-training. Adapt the policy to one task with fine-tuning, human corrections, and reinforcement learning. The LIBERO and RoboTwin checkpoints complete this stage on their benchmarks.

Evaluation

For results after task post-training, see the LIBERO and RoboTwin 2.0 cards.

Uses

Intended uses

Fine-tuning on your own demonstrations, for your own robot and task.
Research on robot learning and world action models.

Out of scope

Deployment without fine-tuning. This checkpoint is not tuned to a specific robot or task.
Use on a physical robot without a risk assessment, an emergency stop, and supervision.
Any use that the license does not permit. This includes weapons, unlawful surveillance, and use in sanctioned countries and regions.

Limitations

Performance on a new robot depends on the fine-tuning data. Collect demonstrations in the robot's own embodiment and action space.
The model can fail on objects, scenes, and instructions that are far from its training data.
This release contains the weights and a minimal inference package. It has no training or fine-tuning code. To fine-tune V0.5 on your own robot and tasks, sign up for the Verne fine-tuning platform at app.vernerobotics.com.

Physical safety

V0.5 outputs actions that move physical machines. A machine that moves can injure people and damage property. Before you run V0.5 on a robot, assess the risks, install an emergency stop and speed and force limits, and supervise the robot while it runs. Section 6 of the license makes you responsible for these measures.

License

The Verne V0.5 Model License governs the use of these weights. In short:

For commercial use, please contact partnership@vernerobotics.com. We would love to work with you!
Use in a country or region under comprehensive US sanctions, and use by a person on a US restricted-party list, is not permitted.
Weapons and unlawful surveillance are not permitted.

This summary is not the license. The LICENSE file controls. See NOTICE for third-party notices.

Citation

@misc{verne2026v05,
  title  = {Verne V0.5: An Open-Weight World Action Model to Accelerate Industrial Robot Deployments},
  author = {{Verne Team}},
  year   = {2026},
  url    = {https://vernerobotics.com}
}