How many parameters does Qwen3.8-Max have and how many are active per query?
Qwen3.8-Max has 2.4 trillion total parameters but uses a mixture-of-experts architecture that activates only around 95 billion parameters for any given query. This sparse activation design dramatically reduces the computing cost and response latency compared to running the full 2.4 trillion parameters on every request, which is how it stays economically viable at scale.
Answered in
Alibaba's 2.4 Trillion-Parameter Model Lands as China's AI Price War Heats UpAlibaba unveiled Qwen3.8-Max, a 2.4-trillion-parameter model, escalating China's five-way AI race alongside DeepSeek, Kimi, Doubao, and ERNIE.
Read the full analysisOther questions this article answers
More ai business & money questions
- What did Oracle announce for Fusion Applications?
- Are the new Fusion agents generally available?
- Who can apply for the Rosalind Academic Plan?
- Does the Rosalind Academic Plan include API credits?
- How many people and for how long does the plan cover?
- What is "AI washing" in the context of corporate layoffs?
- How much has AI language increased on corporate earnings calls?
- Has AI become the leading reason cited for US job cuts?
Every answer on Crashtech is written by the editor of the article it comes from — never auto-summarised. Browse all answers or the AI Business & Money beat.