发布节奏
最近 12 周每周中位浏览
最近 12 周发布时间
颜色 = 中位浏览爆款 vs 日常
爆款 = 浏览前 10%| 指标 | 爆款 | 日常 |
|---|---|---|
| 帖子 | 1 | 7 |
| 带媒体 | 100% | 42.9% |
| 回复 | 0% | 0% |
| 平均字数 | 329 | 273 |
| 中位浏览 | 768.6万 | 2.2万 |
打法建议
只陈述相关性- 一周里 周五 的中位浏览最高。
- 形式上 带图/视频 的中位浏览更高（210.8万 vs 1.9万）。
- 爆款里带媒体占 100%、回复占 0%；日常分别是 42.9%、0%。
- 当前节奏约 0.5 条/天，日均浏览 150万。
内容结构
按本站分析的分类（近 120 天）还没有分类分析。
形式与语言
帖子占比- 纯文字4
- 图片3
- 视频1
回复占比 0% · 引用 50% · 带媒体 50%
互动最高的帖子
按浏览排序- 1
Qwen-Audio-3.0-TTS is here. 🎙️ Our latest text-to-speech model, in two flavors: • Flash: real-time interaction • Plus: high-quality generation What's new: • Multilingual coverage across 16 languages • Style control in natural language • Fine-grained tags for non-verbal details • More robust voice cloning from imperfect audio
- 2
The Qwen model family cyber atlas: a visual walkthrough of how AI really evolved — from large language models to embodied intelligence. It all begins with large language models. Watch the full video below.
- 3
Qwen-Audio-3.0-TTS-Plus is now #1 on the Artificial Analysis Speech Arena.🎙️
- 4
Hello, creators and builders, This week was packed with major drops: Qwen3.8 (2.4T parameters) is coming open-weight, with Qwen3.8-Max-Preview now available on Alibaba’s Token Plan. Qwen-Image-3.0 brings a real productivity tool with rich content, authentic details, and deep knowledge. Qwen-Audio-3.0-TTS launches with production-grade speech synthesis across 16 languages. And Zvec 0.6.0 is live with faster retrieval and simpler deployment. Let’s dive in. https://open.substack.com/pub/tongyilab/p/july-24-2026-the-tongyi-weekly?r=6te3ta&utm_campaign=post&utm_medium=web&showWelcomeOnShare=true
- 5
The next evolution of AI image generation is here! 🎨✨ Qwen-Image-3.0 is all about one word: "Real". Moving beyond just pretty pictures, it now handles 4.5k token prompts, complex multi-grid layouts, 10px legible text, and even full LaTeX pages. A true productivity powerhouse for design, education, and e-commerce. Go create. 🎨
- 6
😎😎😎
- 7
Zvec 0.6.0 is live. New Group-By search adds per-group Top-K. Random rotation boosts INT4 recall by 50+ points with QPS unchanged. FTS now supports 34+ languages via the UAX #29 standard tokenizer. Better recall, smarter retrieval, broader language support.
- 8
Hello, creators and builders, This week, Wan-Streamer v0.2 brings real-time duplex video conversation with ~550ms end-to-end latency. New feature — music to dance — just landed on http://wan.video. And QwenPaw also shipped a preview of their new terminal UI agent in QwenPaw 2.0. Let's dive in. https://open.substack.com/pub/tongyilab/p/july-10-2026-the-tongyi-weekly?r=6te3ta&utm_campaign=post&utm_medium=web&showWelcomeOnShare=true
点赞、转发、浏览是采集那一刻的快照，不是帖子最终数据；时间为北京时间。KPI 卡片里的同比只在账号有 8 周以上数据时显示， 否则只看近 4 周。“打法建议”只说明数据里的相关性，不代表因果。