WEBVTT

1
00:00:00.260 --> 00:00:05.200
Imagine four cells in a row and a token that must reach the last
cell.
设想四个横向格子，标记需要到达最后一个格子。

2
00:00:06.000 --> 00:00:11.060
The rule is simple: each move must go to the adjacent cell,
without skipping positions.
规则很简单：每步只移到相邻格子，不能跳过位置。

3
00:00:12.340 --> 00:00:18.200
One sequence visits every intermediate cell; another jumps
directly from the start to the final cell.
一个序列经过所有中间格子，另一个直接跳到终点。

4
00:00:19.040 --> 00:00:23.020
They finish in the same cell, but only one route respects the
stated rule.
它们停在同一格，但只有一条路径遵守规则。

5
00:00:24.000 --> 00:00:28.740
These are original schematic sequences, not clips generated by
the models evaluated in the paper.
这是原创示意序列，不是论文评估模型生成的片段。

6
00:00:29.720 --> 00:00:34.960
They show why a final picture alone cannot establish that a
visual procedure was valid.
它们说明最终画面不能单独证明视觉过程有效。

7
00:00:35.587 --> 00:00:43.467
VGI-Bench studies visual reasoning expressed through generated
video, with tasks that require valid intermediate states.
VGI-Bench 研究生成视频表达的视觉推理，要求中间状态有效。

8
00:00:44.287 --> 00:00:49.207
Each model receives an input image as the first frame and a text
description of the task.
模型接收作为首帧的输入图像，以及任务文字描述。

9
00:00:49.787 --> 00:00:56.447
The instructions define the goal, relevant objects, allowed
actions, and shortcuts that must not be taken.
指令规定目标、相关物体、允许操作和禁止的捷径。

10
00:00:57.327 --> 00:01:02.507
The generated clip must make progress toward that goal while
preserving the task's rules.
生成片段需朝目标推进，同时保留任务规则。

11
00:01:03.227 --> 00:01:09.727
A motionless clip can avoid many violations yet accomplish
little; a jump can reach the endpoint incorrectly.
静止片段可能避免违规，却没有进展；跳跃可能错误地到达终点。

12
00:01:10.467 --> 00:01:16.227
The benchmark therefore measures both goal completion and the
process that produced the apparent solution.
因此基准同时衡量目标完成和产生表面答案的过程。

13
00:01:16.713 --> 00:01:21.573
The suite organizes twenty-seven task types into four mutually
exclusive domains,
套件把二十七种任务归入四个互斥领域，

14
00:01:22.033 --> 00:01:23.973
according to their visual characteristics.
依据的是各任务的视觉特点。

15
00:01:24.833 --> 00:01:28.313
Visual organization concerns arranging or selecting objects,
视觉组织涉及排列或选择物体，

16
00:01:28.753 --> 00:01:32.833
while physical manipulation involves plausible object-level
interaction.
物理操控涉及合理的物体交互。

17
00:01:33.513 --> 00:01:40.733
Structured puzzles require explicit rule-following, and
spatiotemporal dynamics concern how states depend on time.
结构化谜题要求遵守明确规则，时空动态关注状态随时间的依赖。

18
00:01:41.693 --> 00:01:44.313
Seven nonexclusive skill tags give another view,
七种可重叠技能标签提供另一视角，

19
00:01:44.613 --> 00:01:49.473
including planning, spatial reasoning, temporal reasoning, and
topology.
包括规划、空间推理、时间推理和拓扑。

20
00:01:50.153 --> 00:01:54.013
One task belongs to one domain but can require several skills at
once.
一个任务属于一个领域，却可能同时需要多种技能。

21
00:01:54.873 --> 00:02:01.353
That separation helps diagnose whether a failure concerns the
visual task family or a particular underlying skill.
这种区分帮助判断失败来自任务类别还是某项底层技能。

22
00:02:02.020 --> 00:02:08.400
Tasks appear at three difficulty levels, with prompts and
initial images reviewed for the intended procedure.
任务有三个难度等级，提示与初始图像按预期过程审核。

23
00:02:09.240 --> 00:02:15.160
The benchmark favors photorealistic-style inputs to reduce
mismatch with video models' learned visual appearances.
基准采用照片真实感输入，减少与模型已学视觉外观的不匹配。

24
00:02:16.140 --> 00:02:20.920
Before admitting a task, a small easiest-level sample is tried
on several video models.
接纳任务前，会让多个视频模型尝试少量最简单样本。

25
00:02:21.520 --> 00:02:24.860
At least one tested model must solve it, and at least one must
fail.
至少一个被测模型需成功，同时至少一个需失败。

26
00:02:26.040 --> 00:02:32.740
This screen avoids trivially solved or wholly infeasible
proposals; it does not certify every instance's solvability.
这排除过于简单或完全不可行的方案，不认证每例都可解。

27
00:02:33.840 --> 00:02:36.800
Our abstract four-cell drawing explains the rule distinction;
抽象四格图只解释规则区别；

28
00:02:37.340 --> 00:02:39.980
it does not represent the benchmark's realistic input style.
它不代表基准的真实感输入风格。

29
00:02:40.427 --> 00:02:46.967
Completeness asks how far the video progresses, using
task-specific complete, partial, or failed tiers.
完成度用任务特定的完成、部分完成或失败等级衡量进展。

30
00:02:47.707 --> 00:02:52.547
The tiers map to one, one-half, and zero in the reported
evaluation formulation.
这些等级在评估中映射为一、二分之一和零。

31
00:02:53.587 --> 00:02:59.187
A separate rubric checks local rules and counts violations,
using focused inspection of relevant intervals.
独立规则表检查局部规则并计数违规，重点查看相关区间。

32
00:03:00.047 --> 00:03:05.947
An item with more violations receives a smaller score; the
rubric averages those item-level values.
违规越多，条目分数越低；规则分数再对条目求平均。

33
00:03:07.007 --> 00:03:12.967
The final score multiplies completeness by rubric adherence, so
either kind of failure reduces the result.
最终分将完成度与规则遵守相乘，任一失败都会降低结果。

34
00:03:13.847 --> 00:03:19.527
This composite is different from strict success, which
separately requires perfect completion and rubric scores.
该综合分不同于严格成功率，后者要求两项都完全正确。

35
00:03:20.147 --> 00:03:25.867
The main evaluation uses half the instances of each task,
selected with a fixed random seed.
主要评估使用每个任务一半实例，按固定随机种子选择。

36
00:03:26.507 --> 00:03:31.747
A Gemini three Flash evaluator applies the completion and rubric
procedure to generated clips.
Gemini 3 Flash 评估器对片段应用完成度与规则评价。

37
00:03:32.667 --> 00:03:37.687
Seedance two point zero reports the highest overall score in the
table, fifty-one point zero.
表中 Seedance 2.0 的总体分最高，为五十一点零。

38
00:03:38.527 --> 00:03:44.667
Selected other rows report forty-four point zero for Kling three
and thirty-six point seven for Sora two.
所选其他行中，Kling 3 为四十四点零，Sora 2 为三十六点七。

39
00:03:45.767 --> 00:03:49.647
These are composite benchmark scores under published generation
configurations,
这些是公开生成配置下的综合基准分，

40
00:03:50.007 --> 00:03:53.067
not universal percentages of visual reasoning correctness.
不是通用视觉推理正确率。

41
00:03:54.087 --> 00:03:57.847
An image-output adaptation measures static target rendering
separately,
图像输出改编单独衡量静态目标呈现，

42
00:03:58.207 --> 00:04:00.667
without certifying the intermediate video process.
不认证视频中间过程。

43
00:04:01.267 --> 00:04:05.927
Another analysis studies decoded intermediate products during
denoising,
另一分析研究去噪中解码的中间产物，

44
00:04:06.307 --> 00:04:08.747
rather than events along the finished video's timeline.
而非成片时间线中的事件。

45
00:04:09.587 --> 00:04:12.967
The experiment samples one hundred seventeen task instances
实验抽样一百一十七个任务实例，

46
00:04:13.367 --> 00:04:16.927
and compares successive decoded checkpoints from tested open
models.
比较被测开放模型的相邻解码检查点。

47
00:04:17.787 --> 00:04:24.807
A stable state stays the same; self-correction means an
identifiable wrong state becomes a correct state.
稳定状态保持不变；自我纠错指可辨认的错误状态变为正确。

48
00:04:25.627 --> 00:04:29.867
Changing from one wrong solution to another wrong solution does
not count as correction.
从一种错误答案变为另一种错误答案，并不算纠错。

49
00:04:31.067 --> 00:04:37.987
In this sampled analysis, wrong-to-correct transitions remain at
or below zero point nine percent per checkpoint pair.
这一抽样分析中，每对检查点的错到对转移不超过百分之零点九。

50
00:04:38.927 --> 00:04:44.867
The finding describes these decoded-state experiments; it does
not prove that all video models lack internal reasoning.
该发现描述这些解码状态实验，不证明所有视频模型都缺乏内部推理。

51
00:04:45.453 --> 00:04:52.193
Return to the token sequences: both finish in the same cell, but
one skips a required intermediate state.
回到标记序列：都停在同一格，但一个跳过了必要中间状态。

52
00:04:53.053 --> 00:04:57.773
That distinction remains visible only when we inspect the
process, not just the last frame.
只有检查过程，这个区别才可见，不能只看最后一帧。

53
00:04:58.793 --> 00:05:05.133
A useful generated procedure must make progress while keeping
the relevant rules and object states coherent.
有用的生成过程需推进目标，并保持相关规则与物体状态一致。

54
00:05:06.033 --> 00:05:11.693
VGI-Bench makes those two obligations explicit through task
design and its paired evaluation criteria.
VGI-Bench 通过任务设计和成对评价指标明确这两项要求。

55
00:05:12.453 --> 00:05:18.853
The goal is a diagnostic view of visual intelligence, rather
than mistaking visual polish for a valid solution.
目标是诊断视觉智能，避免把画面精致误当成有效解答。

56
00:05:19.813 --> 00:05:24.313
Ask two questions of the clip: did it reach the goal, and was
its route allowed?
观看片段时问两件事：是否到达目标，以及路径是否被允许？