Get the verdict
Same-prompt test 12 · video · October 2026

Concert crowd with confetti: which AI does it best?

Veo 3.1 Fast wins with 7.5/10. Behind it: Kling 3.0 (7.3), PixVerse V6 (7.3), Runway Gen-4.5 (6.8). One 4-second clip per model, unedited, with whatever audio the model made.

The promptA crowd at a concert, hands up, confetti falling, lead singer on stage.
#ModelOveralladherencemotionqualityartifactsCostRender
1Veo 3.1 Fast7.59786$0.6042s
2Kling 3.07.37787$0.5289s
3PixVerse V67.37787$0.2146s
4Runway Gen-4.56.87776$0.4883s

Best per category: adherence Veo (9) · motion all tied (7) · quality Veo & Kling & PixVerse (8) · artifacts Kling & PixVerse (7)

Every output, ranked

#1 Veo 3.1 Fast

Veo 3.1 Fast 7.5
4s · audio · ~60 cr ($0.60) · 42s render

adherence 9motion 7quality 8artifacts 6
Hands up, heavy confetti, singer on stage; hands mostly coherent.

#2 Kling 3.0

Kling 3.0 7.3
4s · audio · ~52 cr ($0.52) · 89s render

adherence 7motion 7quality 8artifacts 7
Singer-focused, confetti late, crowd small in frame.

#3 PixVerse V6

PixVerse V6 7.3
4s · audio · 48 PixVerse cr ($0.21) · 46s render

adherence 7motion 7quality 8artifacts 7
Singer-focused with quick cuts to a close-up; hands up, but little confetti.

#4 Runway Gen-4.5

Runway Gen-4.5 6.8
4s · ~48 cr ($0.48) · 83s render

adherence 7motion 7quality 7artifacts 6
Hands up, smoke cannons, singer; no confetti visible.

Try Google Veo ↗

⚖️
Outputs are shown exactly as the API returned them (4s, 720p, default settings). Scores are editorial judgments on a fixed rubric, given without knowing which model made each output. One run is a snapshot, not a guarantee. Methodology

FAQ

Which AI is best for "AI video of a crowd scene"?

In our October 2026 test of this exact prompt, Veo 3.1 Fast wins with 7.5/10, ahead of Kling 3.0 (7.3), PixVerse V6 (7.3), Runway Gen-4.5 (6.8). Veo 3.1 Fast's output: Hands up, heavy confetti, singer on stage; hands mostly coherent.

What exact prompt was used?

“A crowd at a concert, hands up, confetti falling, lead singer on stage.” Every model got the same prompt with default settings, 4 seconds at 720p, and one best-effort retry only if a generation failed.

How were the outputs scored?

Each output was scored without knowing which model made it, from 1 to 10 on adherence, motion, quality, artifacts. The overall score is the average of those.