architectures · qwen

Qwen3.5 Architecture (Scaled MoE)

Alibaba Qwen3.5 extended MoE architecture. Builds on Qwen3 with wider expert banks (up to 397B total / 17B active), Grouped-Query Attention, and dynamic expert dropout during training. Supports 262K context via dual ROPE scaling. Used by Qwen3.5-27B through Qwen3.5-397B-A17B leaderboard models.

Verify

What this registry has checked

Quarantine. We hold the publisher’s own published hash for these weights and have recorded it against this entry. We have not downloaded and hashed the weights ourselves, so the hash is the publisher’s claim, not our measurement.
Weight file SHA-256bdb2545734738da2dcd1b68c0ae9a896b4823206efe129fb48500eefe53dbd35
Re-check this yourself → Source → Full record →

What is recorded

Publisherqwen
OriginRecorded as a covered nation under 10 U.S.C. § 4872(f) — the PRC, Russia, Iran or North Korea. The registry did not record which.
LicenceApache-2.0
Pulls recorded here92,466

Elsewhere in the registry

All architectures → Browse with filters → What will run on my hardware? →

Registry record as of 2026-08-14. Identity, licence and parameter count are properties of a release and do not move; tier can, and the verify link above re-checks it against the live registry rather than this page.