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.

Paper
Data
Model
EvalKit
Leaderboard

Data Engines

View All
circle_central_dot
GitHub
Knowledge out-of-domain testset
Prompt
A row of dots is shown. Circle the dot that is in the middle by count (the one with an equal number of dots on each side).
First Frame
First Frame
Last Frame
Last Frame
Video
predict_next_color
GitHub
Abstraction in-domain testset
Prompt
Predict the next color in the sequence.
First Frame
First Frame
Last Frame
Last Frame
Video
maze
GitHub
Spatiality out-of-domain testset
Prompt
The scene shows a 15×15 grid maze with dark walls and white pathways. A green circular marker indicates the starting position, and a red flag marks the end position. Starting from the green start position, navigate through the maze by moving along the white pathways. You can move to adjacent cells (up, down, left, right) but cannot pass through the dark walls. The goal is to find and demonstrate the complete path from the green start to the red flag end position, showing each step of the journey through the maze.
First Frame
First Frame
Last Frame
Last Frame
Video
symbol_edit
GitHub
Transformation out-of-domain testset
Prompt
The sequence currently has 1 of symbol △. Constraint: at least 4 of symbol △. Insert 3 △ symbols at positions 1, 3, and 7 to satisfy the constraint.
First Frame
First Frame
Last Frame
Last Frame
Video
find_fragment_for_gap_filling
GitHub
Perception training set
Prompt
The scene has two separated areas: a top PUZZLE area and a bottom CHOICES area. In the PUZZLE area, the center shape has a missing cut-out outlined in black. In the CHOICES area, there are 4 candidate pieces of the same shape but different sizes. First compare the candidate sizes and determine which single option would fit the missing cut-out exactly (a perfect match in size). Do not use color as a clue. Then circle the correct option. Show the complete solution step by step.
First Frame
First Frame
Last Frame
Last Frame
Video

Inference Results

View Full Bench
Circle Central Dot - Samples
00
01
02
03
04
Task Domains 1/5
Circle Central Dot
Knowledge out-of-domain testset
Shape Color Then Move
Abstraction out-of-domain testset
LEGO Construction
Spatiality in-domain testset
Rotation Puzzle
Transformation in-domain testset
Arrange By Circumference
Perception out-of-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