Skip to content
VECTOR WIREAI INTELLIGENCE
⌘K
UTC
Models785Deals30d265Policy30d67, 14 pendingSources1,086
Home
Sunday, July 19, 20261 min read← AI WIRE

Alibaba Previews Qwen3.8 with 2.4 Trillion Parameters

Alibaba previewed Qwen3.8-Max-Preview, a 2.4 trillion-parameter model it claims is second only to Anthropic's Claude Fable 5.

By Ed. U. AhmadPublished · 1 min read
Sourced from 17 outletsAdd to Preferred Sources
BY THE AI DESKED. U. AHMAD
BRIEFIndustrySOURCED FROM 17 OUTLETS

Alibaba Group Holding has previewed Qwen3.8, its next-generation AI model, which the company claims is "second only" to Anthropic's Claude Fable 51.

Qwen3.8-Max-Preview, the preview version of the strongest model in the Qwen family, features 2.4 trillion parameters. Alibaba described it as "one of the most powerful" models available.

The model has been made available on Alibaba's Token Plan subscription service, as well as its Qoder and QoderWork agentic platforms, according to an official post by the company on X on Sunday.

Bloomberg separately reported on the unveiling of the flagship model preview2.

The 15 articles this piece was written from — of 17 independent outlets covering the story.Also consulted3Alibaba launches Qwen3.8 with 2.4 trillion parameterstechnode.comMedium4Alibaba outlines Qwen3.8 with 2.4 trillion parameterstechnode.comMedium5Alibaba Ships Qwen App Update and Puts Wan 3.0 Into Public Test, With One Eye on Enterprise Buyerspandaily.comMedium6Alibaba Launches Qwen3.8 Foundation Model and Qianwen Office Agent, Reseating Itself in the Enterprise AI Racepandaily.comMedium7Alibaba Releases Qwen3.8-Max Official Version: 2.4 Trillion Parameters With 1M Context Window, Long-Run Agent Capability, and Open Source Coming Within a Weekpandaily.comMedium8Alibaba Qwen3.8 Open Sources 2.4 Trillion Parameters: Claiming Strongest Model After Fable 5 as Chinese Open-Source Models Surgepandaily.comMedium9Alibaba open-sources 2.4-trillion-parameter Qwen3.8 as Chinese mega-model technical approaches converge - finance.biggo.comnews.google.comMedium10Alibaba debuts Qwen3.8-Max model with 2.4T parameterssiliconangle.comMedium11Alibaba previews Qwen3.8, claims it’s second only to Claude Fable 5siliconangle.comMedium12Alibaba’s new model promises Opus 4.6-level performance on your laptopthenewstack.ioMedium13Alibaba releases weights for Qwen3.8 models under Apache 2.0 license, including Qwen3.8-27B, which it says beats Qwen3.7-Plus and excels in real-world coding (@alibaba_qwen)techmeme.comMedium14Qwen3.8-Max: Alibaba's new frontier reasoning modelwww.baseten.coMedium15Qwen 3.8 2.4T is out , no 27b today RIP.reddit.comMedium

ANALYSIS The 2.4-trillion-parameter scale and Alibaba's self-reported positioning just below Anthropic's Claude Fable 5 frame Qwen3.8 as a direct bid for frontier-model status. Distribution across Alibaba's agentic platforms — Qoder and QoderWork — indicates the company is coupling model releases with its own developer and enterprise tooling rather than relying solely on third-party API access.

Was this useful?

The day’s vector, in your inbox.

Vector Wire Daily: one dispatch each publishing day.

More from AI Wire
IndustryJul 20Chinese AI Labs Shift From Price War to Performance RaceIndustryJul 20Chinese Open-Weight AI Models Surge as Businesses Shift From US ProvidersIndustryJul 19Kimi K3 and the Chip Selloff Collide, Testing the AI Capex ThesisIndustryJul 20Anthropic Launches Claude Fable 5 With Safety Redirect to Opus 4.8