FrontierThe story, in brief

Alibaba Previews Qwen3.8-Max, a 2.4 Trillion-Parameter Multimodal Model, Days After Moonshot’s Kimi K3 Open-Weight Launch

2.4 trillion parameters. Alibaba just previewed Qwen3.8-Max—and it's claiming second-place only to Fable 5.

Illustration of a transparent lens revealing connected networks across layers of paper.
Exploring the next frontier of AI research.AI illustration by KeyNews
The KeyNews take

Why it matters

Alibaba is escalating the frontier model race with a massive multimodal MoE release, but key benchmarks and pricing details are suspiciously absent. This is a capability claim without proof—investors and builders need to watch what actually ships.

The key facts

6 to know
  1. Qwen3.8-Max-Preview: 2.4 trillion parameters, multimodal MoE architecture

  2. Alibaba claims 'second only to Fable 5' in capability

  3. Preview live on Token Plan, Coder, CoderWork at 10% standard pricing

  4. Missing: benchmark tables, model card, license, per-token pricing, active-parameter count

  5. Timing: Days after Moonshot's Kimi K3 open-weight launch

  6. UNVERIFIED: No independent benchmarks provided; claims lack supporting data

Go to the source

MarkTechPostmarktechpost.com

Publisher excerpt: Alibaba's Qwen team previewed Qwen3.8-Max-Preview, a 2.4 trillion-parameter multimodal MoE model it calls "second only to Fable 5." The preview is live on Token Plan, Qoder, and QoderWork at 10% of standard pricing. What is not live: any benchmark table, model card, license, per-token price, or…
Read original report
Back to today's editionMore frontier news

Keep reading

Related stories

More from Frontier