Arena Image-to-Video
Higher is better · Image-input models only
Users upload an image with a prompt; two anonymous models turn it into a video, and the user votes for the one they prefer. Scores are Elo ratings from head-to-head human votes. Higher is better.
Top score1495H3
Models tested8
Last updated2026.09.22
Model release date (newest on the right)
Best score so farHigher on the chart is better
Which AI makes the images people prefer most?AA Text-to-Image
GPT Image 2.5 Sunburst🥇GPT Image 2.5 Sunburst🥇Which AI edits images the way people prefer most?AA Image EditGPT Image 2.5 Sunburst🥇GPT Image 2.5 Sunburst🥇Which AI makes the videos people prefer most?AA Text-to-Video
H3🥇H3🥇
GPT Image 2.5 Sunburst🥇GPT Image 2.5 Sunburst🥇Which AI edits images the way people prefer most?AA Image EditGPT Image 2.5 Sunburst🥇GPT Image 2.5 Sunburst🥇Which AI makes the videos people prefer most?AA Text-to-VideoH3🥇H3🥇- 🥇MiniMaxH31495
- 🥈GoogleGemini Omni 1.1 Flash1488
- 🥉ByteDanceSeedance 2.51477
- #4ByteDanceSeedance 2.01475
- #5AlibabaHappyHorse 1.01442
- #6xAIGrok Imagine Video1415
- #7GoogleVeo 3.11398
- #8GoogleVeo 3.1 Lite1385
Half of models ≤ 1458Last updated 2026-10-07