Alibaba Cloud·Qwen

Qwen3-Next-80B-A3B

An 80B-total, 3B-active architecture preview combining hybrid linear attention and ultra-sparse MoE.

LegacyLanguageReasoningCodingAgentOpen weights

MODEL NUMBER069

Specifications

Only published specifications are shown.

Total parameters

80B

Active parameters

Architecture

Hybrid

Context

262K

Input modalities

Text

Open weights

Yes

License

Apache-2.0

Status

Legacy

An 80B-total, 3B-active architecture preview combining hybrid linear attention and ultra-sparse MoE.

PreviousQwen3-235B-A22B235B · 128K ctx

CurrentQwen3-Next-80B-A3B80B · 262K ctx

New capabilities Long context

Gated DeltaNetGated Attention

Hybrid reasoningCodingTool useLong context

2025-09-11Open weights

Qwen3-Next-80B-A3B released open weights.

Select a card to view the related model.