---
format: "aidr-story-markdown/v1"
id: "74d28d030872d9108ee57755b3e49fe2784c3fe5cdb5156789b7c76a9ba0a379"
canonical_url: "https://aidr.today/74d28d03?lang=en"
title: "QwenDevs Returns Qwen Image 2.1 Weights to Open Source"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-09-17T18:56:59.000Z"
category: "Models"
topics: ["qwen","open-source","vision-language","fine-tuning","image-21","early-access","vision-model","alibaba"]
source_urls: ["https://marketbrief.now/ai/qwendevs-returns-qwen-image-21-weights-to-open-source-88c9ff6d","https://huggingnews.com/ai/qwendevs-returns-qwen-image-21-weights-to-open-source-88c9ff6d","https://the-decoder.com/alibabas-open-weight-qwen-image-2-1-claims-to-beat-closed-models-in-image-generation-with-just-7-billion-parameters/"]
summary: "A limited group of 50 testers can now apply for early access to the vision model before its general availability. These spots allow a small number of users to…"
---

# QwenDevs Returns Qwen Image 2\.1 Weights to Open Source

> [Open the canonical story](<https://aidr.today/74d28d03?lang=en>)

**Published:** 2026-09-17T18:56:59.000Z
**Category:** Models
**Topics:** qwen, open\-source, vision\-language, fine\-tuning, image\-21, early\-access, vision\-model, alibaba

## Summary

A limited group of 50 testers can now apply for early access to the vision model before its general availability\. These spots allow a small number of users to…

## Sources

- [Story source](<https://marketbrief.now/ai/qwendevs-returns-qwen-image-21-weights-to-open-source-88c9ff6d>)
- [Story source](<https://huggingnews.com/ai/qwendevs-returns-qwen-image-21-weights-to-open-source-88c9ff6d>)
- [Story source](<https://the-decoder.com/alibabas-open-weight-qwen-image-2-1-claims-to-beat-closed-models-in-image-generation-with-just-7-billion-parameters/>)