Skip to content
News on crypto, fintech and digital infrastructure

Search results for «MoE Architecture»

  1. Qwen4-Max Alibaba Debuts as 2.4 Trillion Parameter MoE Model via API

    Alibaba launched Qwen4-Max on November 16, 2026, as a 2.4 trillion parameter sparse MoE model accessible exclusively through its cloud platforms.

  2. Qwen4-Max Alibaba Debuts as 2.4 Trillion Parameter MoE Model with 1M Token Context

    Alibaba made Qwen4-Max generally available on November 16, 2026, introducing a sparse MoE model with 2.4 trillion parameters and a 1 million token context window.