# Qwen 3.8

> **TL;DR.** Qwen3.

- **Category:** AI / Foundation Models / Open-Source LLMs
- **Stage:** emergent
- **Age:** 8 days
- **Origin date:** 2026-07-19
- **First detected:** 2026-07-20
- **Canonical URL:** https://earlyterms.com/term/qwen-3-8
- **Sources:** 6 primary URLs

## Definition

Qwen3.8, branded Qwen3.8-Max-Preview, is Alibaba's flagship 2.4 trillion parameter multimodal Mixture-of-Experts model, the Qwen team's first release above 1 trillion parameters. It natively processes text, images, video, and documents, and targets coding, full-stack development, and office workflows.

Alibaba previewed it July 19, 2026 at Shanghai's World AI Conference, two days after Moonshot AI shipped the 2.8T-parameter Kimi K3. Qwen's team called the new model "[second only to Fable 5](https://x.com/Alibaba_Qwen/status/2078759124914098291)" — Anthropic's flagship — and promised an open-weight release "soon," a break from Alibaba's pattern of keeping [Max-tier models](https://www.marktechpost.com/2026/07/19/alibaba-previews-qwen3-8-max-a-2-4-trillion-parameter-multimodal-model-days-after-moonshots-kimi-k3-open-weight-launch/) closed.

## Analogy

Like an automaker unveiling a supercar at a show while promising to publish the engine schematics next month.

## Why it's emerging now

Alibaba previewed Qwen3.8-Max on July 19, 2026, two days after Moonshot AI's 2.8T Kimi K3 — a 2.4T-parameter multimodal MoE model the Qwen team calls "second only to Fable 5." No benchmarks are published yet, but the 927-point HN thread and the promised open-weight release turned it into the week's flagship Chinese-AI story.

## Related terms

- *competitor:* Kimi K3
- *related:* Claude Fable 5
- *competitor:* GLM 5.2
- *parent:* Qwen3.6
- *parent:* Qwen3
- *alias:* Qwen
- *competitor:* DeepSeek V4 Pro
- *competitor:* GPT-5.4
- *competitor:* Gemini 3.1 Pro
- *related:* Mixture-of-Experts
- *related:* Moonshot AI
- *related:* Qoder

## Sources

1. [MLQ — Alibaba Launches Qwen 3.8 With 2.4 Trillion Parameters](https://mlq.ai/news/alibaba-launches-qwen-38-with-24-trillion-parameters-claims-near-frontier-performance/)
2. [The Decoder — Qwen Takes On Kimi K3 With Open-Weight Qwen 3.8](https://the-decoder.com/alibabas-qwen-takes-on-kimi-k3-with-open-weight-qwen-3-8-says-model-is-second-only-to-fable-5/)
3. [MarkTechPost — Alibaba Previews Qwen3.8-Max](https://www.marktechpost.com/2026/07/19/alibaba-previews-qwen3-8-max-a-2-4-trillion-parameter-multimodal-model-days-after-moonshots-kimi-k3-open-weight-launch/)
4. [Hacker News — Qwen 3.8 (927 points)](https://news.ycombinator.com/item?id=48966120)
5. [Alibaba_Qwen — launch tweet](https://x.com/Alibaba_Qwen/status/2078759124914098291)
6. [NewsBytes — Alibaba's New AI Model Has 2.4T Parameters](https://www.newsbytesapp.com/news/science/alibaba-s-new-ai-model-has-24t-parameters/story)

---
_Generated by EarlyTerms · https://earlyterms.com/term/qwen-3-8_