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.

105–112 of 214

ground truth

Simple Spinner

model outputs

Gemini 3 Flash Preview

A 0.80T 0.33

Qwen3-VL-8B-Instruct

A 0.78T 0.36

GPT-5.4

A 0.83T 0.29

Claude Sonnet 4.6

A 0.76T 0.21

LLaMA 4 Scout

A 0.74T 0.14

ground truth

Simple Spinner

model outputs

Gemini 3 Flash Preview

A 0.79T 0.28

Qwen3-VL-8B-Instruct

A 0.76T 0.23

GPT-5.4

A 0.78T 0.32

Claude Sonnet 4.6

A 0.78T 0.34

LLaMA 4 Scout

A 0.66T 0.29

ground truth

Simple Spinner

model outputs

Gemini 3 Flash Preview

A 0.85T 0.28

Qwen3-VL-8B-Instruct

A 0.66T 0.25

GPT-5.4

A 0.87T 0.37

Claude Sonnet 4.6

A 0.86T 0.26

LLaMA 4 Scout

A 0.59T 0.26

ground truth

Simple Spinner

model outputs

Gemini 3 Flash Preview

A 0.81T 0.19

Qwen3-VL-8B-Instruct

A 0.76T 0.14

GPT-5.4

A 0.81T 0.31

Claude Sonnet 4.6

A 0.85T 0.27

LLaMA 4 Scout

A 0.78T 0.17

ground truth

Simple Spinner

model outputs

Gemini 3 Flash Preview

A 0.92T 0.30

Qwen3-VL-8B-Instruct

A 0.86T 0.31

GPT-5.4

A 0.96T 0.31

Claude Sonnet 4.6

A 0.93T 0.31

LLaMA 4 Scout

A 0.75T 0.22

ground truth

Simple Spinner

model outputs

Gemini 3 Flash Preview

A 0.88T 0.43

Qwen3-VL-8B-Instruct

A 0.72T 0.15

GPT-5.4

A 0.92T 0.28

Claude Sonnet 4.6

A 0.87T 0.30

LLaMA 4 Scout

A 0.73T 0.26

ground truth

Simple Spinner

model outputs

Gemini 3 Flash Preview

A 0.81T 0.30

Qwen3-VL-8B-Instruct

A 0.67T 0.40

GPT-5.4

A 0.93T 0.32

Claude Sonnet 4.6

A 0.87T 0.33

LLaMA 4 Scout

A 0.67T 0.29

ground truth

Simple Spinner

model outputs

Gemini 3 Flash Preview

A 0.96T 0.38

Qwen3-VL-8B-Instruct

A 0.78T 0.23

GPT-5.4

A 0.97T 0.30

Claude Sonnet 4.6

A 0.97T 0.65

LLaMA 4 Scout

A 0.85T 0.17
← Previous14 / 27
go to
Next →