Will someone make a video generator < 3B parameters as good as Sora by EOY 2025?
6
Ṁ140Ṁ265resolved Mar 18
Resolved
NO1H
6H
1D
1W
1M
ALL
This question is managed and resolved by Manifold.
Market context
Get
1,000 to start trading!
🏅 Top traders
| # | Trader | Total profit |
|---|---|---|
| 1 | Ṁ33 | |
| 2 | Ṁ32 | |
| 3 | Ṁ26 | |
| 4 | Ṁ25 |
Sort by:
bought Ṁ40 NO
What did exist by 2025:
There were sub-3B text-to-video models;
for example Wan2.1 has a 1.3B variant (alongside a larger 14B one).
The open / competitive models often cited as approaching top closed models were generally much larger than 3B (e.g., HunyuanVideo ~13B, Mochi ~10B, Wan2.x 5B/14B).
So unless your bar for “as good as Sora” is defined in a very specific, narrow way (e.g., one benchmark or one style niche), the best defensible answer is:
No
not confirmed / not established that <3B matched Sora by EOY 2025.
People are also trading
Related questions
Will a Sora-generated video receive over 1 billion total views across social media before July 2026?
10% chance
Best Video Generation Model (by company/organization) EOY 2026
3x3x3 Rubik's Cube solved by video generation model by EOY 2026?
14% chance
High Quality AI-Generated Video Games by prompt before 2027?
10% chance
Will real-time text-to-video generation be viable by 2027?
22% chance
High Quality AI-Generated Video Games by prompt before 2028?
15% chance
High Quality AI-Generated Video Games by prompt before 2035?
49% chance
High Quality AI-Generated Video Games by prompt before 2040?
70% chance
High Quality AI-Generated Video Games by prompt before 2045?
73% chance
High Quality AI-Generated Video Games by prompt before 2030?
23% chance