A Very Big Video Reasoning Suite

We bet on a future that video reasoning is the next fundamental intelligence paradigm, after language reasoning, where spatiotemporal embodied world experiences could be more naturally captured.

Data Engines

View All
circle_maximum_value
GitHub
Knowledge training set
There are multiple numbers on the screen, circle the one with the largest value
First Frame
Last Frame
shape_color_change
GitHub
Abstraction training set
The scene shows an analogy A:B :: C:? with two rows of shapes and arrows. On the top row, a unknown heart becomes a unknown heart; treat this as the color rule from unknown to unknown. On the bottom row, the pentagon starts with the same source color. Apply the same color change so the answer shape on the right is a unknown pentagon, keeping its shape, size, and position the same while only the color changes.
First Frame
Last Frame
grid_number_sequence
GitHub
Spatiality in-domain testset
The scene shows a 10x10 grid with a green start point, a red end point, and yellow cells marked with numbers 1, 2, and 3. An orange circular agent is positioned at the green start point. The agent can move to adjacent cells (up, down, left, right). Starting from the green start point, the agent must visit the numbered yellow cells in numerical order (1, then 2, then 3), taking the shortest path between each consecutive pair of numbered cells. The agent is allowed to pass through the red end point when visiting the numbered cells if needed. After visiting all numbered cells in sequence, the agent must reach the red end point, also following the shortest path.
First Frame
Last Frame
combined_objects_spinning
GitHub
Transformation training set
The scene shows 2 objects on the left side and dashed target outlines on the right side. The dashed target outlines remain completely stationary. For each object, first rotate it in place to match the orientation of its corresponding dashed target outline, then move it horizontally to the right so that it aligns exactly with and fits within its corresponding dashed target outline.
First Frame
Last Frame
highlight_horizontal_lines
GitHub
Perception training set
The image contains several straight lines, including horizontal and vertical lines. Circle every horizontal line. Do not circle any vertical line.
First Frame
Last Frame

Inference Results

View Full Bench
Dot to Dot - Samples
00
01
02
03
04
Task Domains 1/5
Dot to Dot
Knowledge in-domain testset
Ball Color
Abstraction in-domain testset
Outline Innermost Square
Spatiality out-of-domain testset
Track Object Movement
Transformation in-domain testset
Stable Sort
Perception in-domain testset
Prompt
Loading...
Ground Truth
First
First Frame
Final
Final Frame
Model Outputs
1/
VBVR-Wan2.2
VBVR-Wan2.2
CogVideoX 1.5
Kling 2.6
LTX-2
Runway Gen-4
Sora 2
Veo 3
Wan 2.2 I2V
Hunyuan I2V
Seedance 2.0

Leaderboard

Modality
Split
Type
Category
2026-04-28