We're currently developing in a closed-source environment to prevent potential misuse and intellectual property theft. As AI competition intensifies, we believe this approach is the most effective safeguard against unethical practices. For instance, companies could leverage our tools to build superior models in a closed-source manner without proper attribution, misleading the public. We are committed to fully open-sourcing once the critical significance of our technology is firmly established.