Qwen-Image 2.1 going open source, 50 early access spots open
cadissimus · reddit · 2026-09-17
The Qwen team announced that Qwen-Image 2.1 will be open source, and is opening 50 early access spots for users to try the new image generation model ahead of its official release.
Related event: Qwen announces Qwen-Image 2.1 will be open-sourced, early access open(3 posts)→
More from Multimodal
- Phosphene 4.14 adds local song writing, one-click face fix, and audio tracks on Macs — cocktailpeanut · 2026-09-18
- Phosphene brings free local YuE2 song generation to Macs with one-click launch — cocktailpeanut · 2026-09-18
- Haruomi Hosono's new music video made with Kling hailed as one of the best AI MVs yet — Merzmensch · 2026-09-18
- Recreating nostalgic Mediterranean and Rome scenes with ChatGPT Image 2.5, prompts included — HeyZoyaKhan · 2026-09-18
- SAM3 vs Astra for text-prompt segmentation: sharper masks vs better language understanding — kate_saenko_ · 2026-09-18
- Three.js r186 ships native 3D Gaussian Splatting: turn a 30-second phone scan into a viewable 3D model — nptacek · 2026-09-18