Navigate Select ESC Close

I Created the Same Animation in Every AI Video Generator

2026-02-01 Science & Technology
28.5k
943
96
Futurepedia
Futurepedia
711.0k subscribers

Unlock all features

FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.

Description

Get all cutting-edge AI image and video models in one place with Artlist: https://artlist.io/artlist-70446/?artlist_aid=futurepedia_io_3995&utm_source=affiliate_p&utm_medium=futurepedia_io_3995&utm_campaign=futurepedia_io_3995 More from Futurepedia: 👉 Join the fastest-growing AI education platform! Try it free and explore 30+ top-rated courses in AI: https://bit.ly/futurepediaSL Summary: I tested the 9 leading AI video generators with the same prompts. I used text to video and image to video across 18 different challenges and ranked them all so you know which one is best for what you need. The models tested were Veo 3.1, Kling 2.6, Sora 2, Grok Imagine, Hailuo 2.3, Wan 2.6, Seedance 1.5, LTX 2.0, and Runway Gen 4.5 Chapters 0:00 Intro 1:06 Basketball shot 5:11 Invasion 9:28 Breakdancing 12:10 Joke 15:30 Crying 17:27 Title screen 18:52 Hello 21:00 Alarm 22:58 Text to video Totals 24:12 Dragon tea 27:12 Rack focus 29:16 Car chase 33:08 Robot piano 35:40 Octopus bartender 41:08 King speech 44:04 Poker (Midjourney style) 46:42 Mixing styles 49:16 New York diversity 53:35 Gladiator vs manticore 56:35 Image to video totals

Top Comments (10)

@jazzbutcher 2026-02-01

These models (Sora, Veo 3.1, Kling 2.6, LTX‑2, Wan, Seedance, etc.) all have totally different prompt best‑practices. Using one set of generic prompts is really a test of how they handle bad prompting, not what they’re actually capable of when you follow each model’s own guidelines.

15 5 replies
@danielterrazas 2026-02-01

I can´t believe he never noticed that most of the tentacles (except for Hailou) morphed into human hands.

12
@6AxisSage 2026-02-01

the crying photo ones are wierd because most of them theyre looking at the back of the photo

7
@Plisko1 2026-02-01

Overall this was helpful and well done and I know it was a lot of work. But I gotta say: some of your analysis was baffling. Perfect bartender tentacles that pour in the wrong drink: "FAIL". Tentacles appearing like a beard out of nowhere and turning into hands and arms: "It did great!!!". Manticore's wings vanish completely and it turns into a lion: "Grok did amazing!!!" Seedance keeps the manticore and hits accurate: It scores lower than Grok. You actually highlighted Grok as standing out in two places (bartender and manticore) where it failed pretty obviously. And you specifically identified the Manticore as the complex monster test then gave Grok a pass for not keeping it accurate. It was confusing.

5
@futurepedia_io 2026-02-01

👉 Join the fastest-growing AI education platform! Try it free and get instant access to 30+ top-rated courses in AI: https://bit.ly/futurepediaSL

3 1 replies
@BybbyCreme 2026-02-02

liked how you designed the Video with the Ranking and all. Was definitely one of your most fun videos. Good job, Kevin!

3
@AMR2442 2026-02-01

2:01 he got that Michael Kidd gilcrest 😂😂😂😂

2
@ShiningBlackVFX 2026-02-02

This is arguably the most useful benchmark video I've seen this year. Seeing Veo 3.1 and Kling 2.6 side-by-side really highlights the strengths of each model for cinematic movement. Thanks for the massive effort! 👏

1
@aijedimedia 2026-02-02

When people are fan boys of a model or anything they make up a win for the product or company.

1
@VinhAnh-cage 2026-04-15

Solid comparison, thanks for putting in the work. For keeping the same characters and style consistent across longer animations, I’ve been using Anistudio lately — its agents make the whole pipeline feel way smoother with way less manual tweaking.

0

Unlock the Data Inside
Turn Videos into Knowledge

  • Get FREE 10/day: transcripts, summaries, chats
  • Chat with videos, export text & PDF
  • $1 free API credit for RAG, chatbots & research

Free forever plan • All features unlocked

App screenshot