ModelsAug 25, 2026 · 21:528 sources

Alibaba Ships 120B Qwen3.8-Flash-Next, First Glimpse of Qwen4 Architecture

From huggingnews.com

AI;DR

Alibaba will release the 120B Qwen3.8-Flash-Next multimodal Mixture-of-Experts system for open-weight download on August 26. The model features 120 billion total parameters and 6 billion active parameters to provide a first look at the upcoming Qwen4 architecture. UnslothAI is preparing day zero support for the release.

Read the sourceSign in to suggest

Sources

+ 6 more
  1. SUPPORT@kimmonismus“Alibaba is giving developers early access to the architectural improvements that will underpin the upcoming model family.”x.com
  2. SUPPORT@jun_song“Qwen3.8-120B/51B/A6B MoE is coming out in 24 hours.”x.com
  3. SUPPORT@danielhanchen“working on @UnslothAI day zero support.”x.com
  4. SUPPORT@thezachmueller“Welcome to the era of ~120B models.”x.com
  5. SUPPORT@kimmonismus“roughly 176B parameters of stored capacity with just 6B activated per token”x.com
  6. SUPPORT@wesroth“the ONLY model in its size class in the top 10”x.com

One ranked edition, every morning

The stories that matter today. Free, cancel anytime.

Telegram @aidr_today