🎉 Zvec 0.6.0 Release is Live! This release brings major upgrades across retrieval, compute optimization, deployment experience, and the quantization framework. Upgrade and give it a try! ✨What's new: • 🔍Retrieval: New Group-By search with per-group Top-K support, covering Flat, HNSW, and HNSW-RaBitQ indexes • 📝Full-Text Search: Introduces a UAX #29 standard tokenizer, UTF-8 support, ASCII folding, and Snowball stemming for 34+ languages—a major upgrade for multilingual retrieval; Conjunction queries are now 22–38% faster with block-max pruning • ⚡ Compute Optimization: INT8/INT4 quantization now supports random rotation, evening out the distribution to reduce variance and quantization error; INT4 recall improves by up to 50+ percentage points while QPS remains essentially unchanged • 💾 Deployment Experience: DiskANN and libaio are now dynamically decoupled via dlopen, removing the libaio build dependency and plugin .so; libaio is automatically detected at runtime, with a graceful fallback when unavailable • 🧩 Quantization Framework: The internal quantization module has been refactored into a pluggable Turbo framework, decoupling quantization logic from index code; future quantizers such as PQ and RaBitQ can be added independently without modifying index internals 📚 Learn more: • 📄 Release Notes: zvec.org/en/blog/2026-07-20-… • 🧭 Roadmap: github.com/alibaba/zvec/issu…

🎨 Meet Qwen-Image-3.0 — the third generation of our foundational image generation model. If 1.0 was about "Precision," and 2.0 added "Variety, Completeness, Beauty & Authenticity," then 3.0 comes down to a single word: Real (实). Three dimensions of "Real": 📰 Rich Content — prompts up to 4.5k tokens. One-pass generation of complex layouts: newspapers, storyboards, exam papers — even a 3×3 infographic grid or picture-in-picture-in-picture UIs. 🔬 Authentic Details — text legible down to 10px, full LaTeX paper pages, pores, hair strands & near-photographic skin texture. 🌏 Deep Knowledge — native rendering in 12 languages, 100+ art styles, realistic UIs (web / games / livestreams), plus world knowledge & live web retrieval. Not just "good-looking" — genuinely useful. Image generation as a real productivity tool for design, content, education & e-commerce. Go create 🏃🎨 💬Qwen Chat: chat.qwen.ai/?inputFeature=t… 📝Blog: qwen.ai/blog?id=qwen-image-3…

Alibaba's Qwen-Audio-3.0-TTS-Plus is the new leading model on the Artificial Analysis Speech Arena Leaderboard for Provider Voices, narrowly surpassing Simba 3.2 and ahead of Gemini 3.1 Flash TTS, and Sonic 3.5 Qwen-Audio-3.0-TTS-Plus is Alibaba's latest Text to Speech model, released with increased naturalness and contextually appropriate intonation. The release continues Alibaba's strong momentum across AI model releases, following recent leading launches in language, image, and video generation. Key takeaways: ➤ Quality: Qwen-Audio-3.0-TTS-Plus has an Elo score of 1,236 (+17/-17) based on 1,305 arena appearances, narrowly ahead of Simba 3.2 at 1,234 (+17/-17), with overlapping confidence intervals, and ahead of Gemini 3.1 Flash TTS at 1,214 and Sonic 3.5 at 1,207. ➤ Throughput speed: The model generates 16 characters per second, below other leading models Simba 3.2 (30.2), Gemini 3.1 Flash TTS (27), and Sonic 3.5 (120) ➤ Pricing: Qwen-Audio-3.0-TTS-Plus is priced at $27.59 per 1M characters via Alibaba Cloud Model Studio, below Sonic 3.5 ($39.00/1M), above Gemini 3.1 Flash TTS ($18.31/1M) and Simba 3.2 ($10.00/1M). See more details and listen to samples below ⬇️

Qwen3.8 is launching and going open-weight soon!🌐 With a massive 2.4T parameters, this model is continuously evolving. We believe it’s one of the most powerful model available today, compatible to leading frontier AI models , second only to Fable 5. You don't have to wait to test it. Just now, the Qwen3.8-Max-Preview made its debut on Alibaba’s Token Plan, Qoder, and QoderWork. Be among the very first to try it out. Can't wait to hear what you build. Stay tuned! 🚀  Token Plan international：qwencloud.com/pricing/token-… China：platform.qianwenai.com/prici…

Hi Friends! This is a preview of our terminal UI agent! It will be part of the upcoming QwenPaw 2.0 release. QwenPaw TUI is simple and keyboard-first way to work with your agents configured in QwenPaw, built on the new foundation of QwenPaw 2.0.

Training Wan 2.1-1.3B to use Qwen3-VL-2B text encoder. Doing 33% text only, 33% VL only, and 33% both. Just pretraining the two text input linear layers for now. It is amazing how quickly a model can adapt to a different text encoder. This is 25,750 steps, BS of 10.

Qwen-Image-Edit-2511でピクセル位置パーフェクト編集ができるようにノード改造をおこなった。 メインノードの改造と補助ノードを1つ付けることで、実質、この問題は解決したんじゃないかな。 Qwen-Image-Edit-2511は、現在エディットモデルが使える画像生成AIで、ローカルでLoRA学習ができる限られたモデルのひとつなので、この「ピクセル位置パーフェクト編集を確実にできる」ようにしておくとさらに応用範囲が広がります。つまりツールが作り放題ということ。 この辺りをどう実装すればいいのかは、AIに調査させればおおよそのヒントはでてくるんだけど、「補助ノードが果たしている機能」までは触れてないのが大抵。これはDiTの仕様に関わる問題の解決なので、そこをクリアすればほぼ間違いなくピクセル位置パーフェクト編集ができるようになる。 このムービーはその上で、Qwen-Image-Edit-2511でイラストから線画推定をさせてみたものなんだけど、なんの学習もしなくても毎回安定した線画が抽出できている。しかもPhotoshopで確認してもズレはまったくないと来ていると、このモデル元々とんでもなく性能が高かったんだなと改めて感心してしまいます。

Doing some tests with @Ali_TongyiLab's Qwen3-VL and groundings. This is a very nice model. It is not blazingly fast on my RTX 3050 with 6GB RAM, but it works really well. I bet I can do some more fun stuff with it.