The Beijing-based firm said Sunday it had experienced “unprecedented compute challenges” and would temporarily focus on ensuring it could serve its existing paid users. The large-language model was trained on 2.8 trillion parameters, making it the largest open-source model ever released.