Fantastic! 🥳 Thanks @lightseekorg for the TokenSpeed day-0 support. The new architecture is fully covered, from GDN + QSA to N-gram embeddi…

Fantastic! 🥳 Thanks @lightseekorg for the TokenSpeed day-0 support. The new architecture is fully covered, from GDN + QSA to N-gram embedding with FP8. TokenSpeed Day 0 Support for @Alibaba_Qwen 3.8 Flash Next. 🔹GDN + Qwen Sparse Attention hybrid architecture 🔹Gated residual connections 🔹N-gram embedding For N-gram embedding, TokenSpeed also supports FP8 precision. As a preview for Qwen 4, we will continue optimizing beyond D…

Source: Qwen (X) | 2026-08-27