Qwen3.8-Flash-Next
Alibaba (Qwen)we follow GLM (GLM-5-Team, 2026) and reuse the top- indices across speculative decoding steps
usedunclearin Qwen3.8-Flash-NextQwen Team
1 more from this model
So the draft decode steps stop running the indexer altogether.
coreunclearin Qwen3.8-Flash-NextRadixArk