animation2code benchmark
For best compatibility, please view this dashboard in a Chrome browser.

Zero-shot video (or image-frame) → code results on the test set, across commercial and open-source models.

Each output is tagged with A = appearance similarity and T = temporal similarity; higher is better for both. Click a video to inspect its code.

8996 of 214

model outputs

GPT-5.4

A 0.83T 0.25

Claude Sonnet 4.6

A 0.65T 0.24

LLaMA 4 Scout

A 0.56T 0.07

model outputs

GPT-5.4

A 0.85T 0.38

Claude Sonnet 4.6

A 0.88T 0.27

LLaMA 4 Scout

A 0.69T 0.21

model outputs

GPT-5.4

A 0.82T 0.22

Claude Sonnet 4.6

A 0.86T 0.25

LLaMA 4 Scout

A 0.54T 0.31

model outputs

GPT-5.4

A 0.90T 0.27

Claude Sonnet 4.6

A 0.82T 0.32

LLaMA 4 Scout

A 0.78T 0.18

model outputs

GPT-5.4

A 0.85T 0.11

Claude Sonnet 4.6

A 0.86T 0.22

LLaMA 4 Scout

A 0.82T 0.19

model outputs

GPT-5.4

A 0.91T 0.21

Claude Sonnet 4.6

A 0.85T 0.18

LLaMA 4 Scout

A 0.68T 0.23

model outputs

GPT-5.4

A 0.89T 0.32

Claude Sonnet 4.6

A 0.83T 0.33

LLaMA 4 Scout

A 0.81T 0.28

model outputs

GPT-5.4

A 0.82T 0.32

Claude Sonnet 4.6

A 0.83T 0.32

LLaMA 4 Scout

A 0.55T 0.27