HF card + qwen.ai blog · 2026-08/09 · interactive model
Qwen3.8-Max
2.4T-parameter MoE, ~95B active. 92 layers in a 23×(3 Gated DeltaNet → 1 Gated Attention) rhythm,
every layer ending in a 512-expert MoE. Click any part — slab, rack, tower or dial — to read what it does.
SELF-DESTRUCT IN 6:00:00
this exhibit is temporary · six hours only
drag · scroll · click a part · ← → to step · esc to close
Exhibit closed
This model was provisioned for six hours and has now been decommissioned.
The container is gone; the architecture page in the wiki remains.