A benchmark based on Openworldlib

[hzeroyuke]

Hi OpenWorldLib team,
Our team has recently completed a benchmark for interactive long-video world models. Here is the link: https://alibaba-damo-academy.github.io/WorldOlympiad/
The generation part of our work is built upon OpenWorldLib, and we sincerely appreciate the help from the OpenWorldLib project. We have also further adapted the LongLive and Rolling Forcing pipelines under the OpenWorldLib framework, and fixed some bugs in other pipelines. We will be submitting several PRs in the near future.
The arXiv version of our paper will be released very soon. If the OpenWorldLib team could offer some promotional support, we would be very grateful.
Once again, thank you for the help from the OpenWorldLib project.
Best regards,
The WorldOlympiad Team

[zengbohan0217]

Congratulations on your outstanding work. We are very glad to see that your benchmark can be merged into OpenWorldLib. At the same time, we will also help promote WorldOlympiad.

[hzeroyuke]

Our paper has been published at https://arxiv.org/abs/2606.11129. We would be very grateful if you could upvote it on Hugging Face: https://huggingface.co/papers/2606.11129

[zengbohan0217]

OK, I will ask others to upvote for your paper~