qwen35-122b-a10b
live
citable URL: https://halobench.com/records/qwen35-122b-a10b/ — this address never moves; the anchor /records/#qwen35-122b-a10b keeps resolving
Qwen3.5-122B-A10B-MTP @ UD-Q4_K_M, single slot · kind model · engine igpu · tier daily-driver
runs on aibeast · config cfg-0002
⌁ current state live · ⌁ days in production 46
Lifecycle — append-only
Displaced the benchmark winner on a deliberate override of the data: the point of the exercise was to find what Warden could become, and the 122B bets on stronger cross-domain reasoning showing up in real proactive work. Later reinforced by the MTP variant (+49% decode) settling the choice. Reference-only in the benchmark — it matched the 35B on reactive speed despite 10B active vs 3B, which is what exposed reasoning-token length rather than decode speed as the latency driver.