Qwen3-Next-80B-A3B-Base

by Alibaba Cloud / Qwen Team

Qwen3-Next-80B-A3B-Base is a model from Alibaba Cloud / Qwen Team, released on September 10, 2025.

Qwen3-Next-80B-A3B-Base is the foundation model in the Qwen3-Next series, featuring revolutionary architectural innovations for ultimate training and inference efficiency. It introduces Hybrid Attention combining Gated DeltaNet (75% layers) and Gated Attention (25% layers) for efficient ultra-long context modeling, Ultra-Sparse MoE with 512 total experts but only 10 routed + 1 shared expert activated (3.7% activation ratio), and native Multi-Token Prediction for faster inference. With 80B total parameters and only ~3B activated per inference step, it achieves performance comparable to Qwen3-32B while using less than 10% training cost and delivering 10x+ throughput for 32K+ contexts. Trained on 15T tokens with training-stability-friendly designs including Zero-Centered RMSNorm and normalized MoE router parameters. Supports 256K context length, extensible to 1M tokens with YaRN scaling.

Key specs

ProviderAlibaba Cloud / Qwen Team
Release date2025-09-10
Licenseopen source

Quick answers

Is Qwen3-Next-80B-A3B-Base open source?

Yes. Qwen3-Next-80B-A3B-Base is an open model released under the open source license.

When was Qwen3-Next-80B-A3B-Base released?

Alibaba Cloud / Qwen Team released Qwen3-Next-80B-A3B-Base on September 10, 2025.

Compare on the AI leaderboard · All AI models

More models from Alibaba Cloud / Qwen Team