NVIDIA Introduces GR00T 2 for Robots Learning From Human Video
NVIDIA Introduces GR00T 2 for Robots Learning From Human Video
On November 9, 2026, NVIDIA introduced GR00T 2 at CoRL in Austin. The model uses a 2-billion-parameter transformer built on Cosmos-Predict2.5-2B and is trained on 402,730 hours of first-person human video, 10,184 hours of handheld-gripper demonstrations, and 23,500 hours of robot teleoperation.
Source: Robotics — Robot learning news discovery · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
Reported numbers
human video hours
402,730 hours
View original evidence
nd the next action sequence. One pretrained checkpoint covers parallel-gripper arms, dexterous hands and a humanoid. Although NVIDIA introduces the model as open, its research page says code will be available in December. Pretraining combines 402,730 hours of first-person human video, 10,184 hours of handheld-gripper demonstrations across 2,897 tasks and 23,500 hours of robot teleoperation. The Human-1M-100K release
Open source S3handheld-gripper demonstration hours
10,184 hours
View original evidence
nd the next action sequence. One pretrained checkpoint covers parallel-gripper arms, dexterous hands and a humanoid. Although NVIDIA introduces the model as open, its research page says code will be available in December. Pretraining combines 402,730 hours of first-person human video, 10,184 hours of handheld-gripper demonstrations across 2,897 tasks and 23,500 hours of robot teleoperation. The Human-1M-100K release
Open source S3robot teleoperation hours
23,500 hours
View original evidence
nd the next action sequence. One pretrained checkpoint covers parallel-gripper arms, dexterous hands and a humanoid. Although NVIDIA introduces the model as open, its research page says code will be available in December. Pretraining combines 402,730 hours of first-person human video, 10,184 hours of handheld-gripper demonstrations across 2,897 tasks and 23,500 hours of robot teleoperation. The Human-1M-100K release
Open source S3
Dataset details reported by the source
- Modalities
- video Open source S3
- paper: not_reported
- code: Reported pending Open source S3
- dataset: not_reported
What remains unknown
Dataset licensing and commercial-use rights are not established here.
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
nd the next action sequence. One pretrained checkpoint covers parallel-gripper arms, dexterous hands and a humanoid. Although NVIDIA introduces the model as open, its research page says code will be available in December. Pretraining combines 402,730 hours of first-person human video, 10,184 hours of handheld-gripper demonstrations across 2,897 tasks and 23,500 hours of robot teleoperation. The Human-1M-100K release
Open source S3
NVIDIA introduced GR00T 2 on November 9, 2026, at the Conference on Robot Learning (CoRL) in Austin, and reported 40% zero-shot success on 50 tasks absent from the target robot’s training data for a model trained partly on human video. GR00T 2 uses a 2-billion-parameter transformer built on Cosmos-Predict2.5-2B. It combines recent camera views, robot state and a written instruction to generate expected future views a
Open source S2
opens 100,000 hours of that human data, including video, language annotations and head and hand motion data, under a CC BY 4.0 license permitting commercial and noncommercial use. Cross-embodiment tests show how skills transfer between bodies and adapt with limited robot data. NVIDIA reports 65% success for filter-paper setup learned from human demonstrations without task-specific teleoperation, and 60% for toy dump-
Open source S4
truck assembly transferred from a handheld gripper to a YAM arm. On the Dexter evaluation, GR00T 2 completed at least one trial on 40 tasks. Average success reaches 71% with one demonstration per task and 95% with five; those five supply under three minutes of robot data per task, followed by an approximately 30-minute adapter fine-tune on one RTX 5090. The technical whitepaper qualifies the five-demonstration result
Open source S5
Implications for data suppliers
RoboSignal interpretation and collection questions, not statements of buyer demand.
- Confirm the required data type and collection setting with the buyer; this source does not establish a complete collection specification.
- Compare the reported units and scope before using these quantities in a budget. Recording hours, sensor-hours and trajectories are different measures.
- Check the exact dataset version, license, modalities and task coverage against the intended use.
- Validate demand and acceptance criteria with a buyer before scaling. Publication, popularity and a research result do not establish a purchase commitment.
Source：Robotics — Robot learning news discovery · unite.ai