Skip to content
RobothecaBeta
Log in

RoboInter-VLM

  • AI model
  • Humanoids and robot learning

RoboInter-VLM is a suite of intermediate representation resources for robotic manipulation, including a large annotated dataset, a VQA benchmark, and VLM planners built on Qwen2.5-VL and…

Made by Intern Robotics, Shanghai AI Laboratory.

Source on GitHub: github.com/InternRobotics/RoboInter (opens in a new tab). Facts read at commit 481acfb (opens in a new tab).

Licenses

Code
MIT (opens in a new tab)

Read by Robotheca from LICENSE (opens in a new tab) (MIT) at commit 481acfb (opens in a new tab), on 16 Oct 2026.

Commercial use
Yes

What it does

  • Provides RoboInter-Data, a dataset of over 230k episodes across 571 scenes and 6 robot arm types, with dense per-frame annotations.
  • Offers RoboInter-VQA, a visual question answering benchmark with 9 spatial and 20 temporal categories, about 2.3M samples.
  • Ships VLM planners fine-tuned on RoboInter-VQA that generate intermediate representations such as traces and bounding boxes as high-level plans.
  • Includes RoboInter-VLA, a plan-then-execute framework with end-to-end and modular variants linked by Flexible Chain-of-Thought.
  • Includes RoboInter-Tool, a SAM2-based GUI for semi-automatic annotation of manipulation videos with human review.
  • Provides a Gradio app for interactive visualization of the dataset annotations.

Written by Robotheca from the project's README (opens in a new tab) at commit 481acfb.

Getting started

Difficulty
Advanced

Maintenance

Status
Maintained: a commit in the last 12 months
Newest commit
14 Feb 2026 on main
Latest release
None published on GitHub

Checked on GitHub on 16 Oct 2026.

Links

Robots that can run it

No robot on Robotheca is linked to it yet. Links come from a robot's robot.yaml, from its dependency files read as text, or from Robotheca's editors.

Robotheca lists AI models as a reference: nothing is hosted or run here, and every fact links to where it was read.