Leon做了个

Silicon proposes. Carbon disposes.

[Leon做了个 @leon_made_this]

一个给三岁女儿做的小试验，首发 V2EX 后冲到了“今日热榜”第一🥇，github 仓库也涨破 400 stars🌟。感慨最深的不是 AI 帮我们加速了什么，而是它真的能帮助到我们，即使是生活中也许是最小的最微不足道的一件事情，也能挖掘出很大的意义😄。很高兴能帮助到许多家长和小朋友，一起共享美好的亲子时光～

[Leon做了个 @leon_made_this]

https://t.co/zdjlXeHje1

[Leon做了个 @leon_made_this]

@vaolenhtamlinh 恭喜呀，越南的朋友👬

[Leon做了个 @leon_made_this]

@yqgyx123 很不错了，也算是一分不少收益吧哈哈哈

[Leon做了个 @leon_made_this]

@meditation198 恭喜恭喜🎉

[Leon做了个 @leon_made_this]

早上好呀，全世界各地的朋友们你们都吃什么早餐？吃完早餐又是活力满满的一天🌞

[Leon做了个 @leon_made_this]

@membervic1506 这两天看到不少越南的朋友👭真好 o(^▽^)o

[Leon做了个 @leon_made_this]

@cubb52532767 恭喜恭喜🎉

[Leon做了个 @leon_made_this]

@cuonggiau2 越南的朋友真的很团结呀

[Leon做了个 @leon_made_this]

@SonicAIWizard cool～

[Leon做了个 @leon_made_this]

@Com0loquieras LMAO，必须把高贵的 A/ 狠狠踩在脚下

[Leon做了个 @leon_made_this]

重点标记一下：
1. 原生多模态 Dense 27B
2. 整体超过 Qwen3.7-Plus，强项是真实编程和办公
3. 原生 262K 上下文，YaRN 可扩到 1M
4. Apache 2.0，定位给本地开发者用
5. 可以调思考深度（告别 thinking 开/关）

粗看一下各方面提升都比较大了，本地跑不错～

[Qwen @Alibaba_Qwen]

We promised open weights for Qwen3.8. Now, time to meet them! 🎉

⚡ Qwen3.8-27B:
- A native multimodal dense model. With just 27B parameters, it outperforms Qwen3.7-Plus overall and shines in real-world coding & office workflows.
- 262K native context, easily extendable to 1M tokens via YaRN.
- Built for builders. Highly efficient, high-quality, and licensed under Apache 2.0.

🚀 The open weights for Qwen3.8-2.4T-A95B (Max-level) have also been released recently.

Whether you're shipping lightweight applications with Qwen3.8-27B locally or building agents with Qwen3.8-2.4T-A95B, they're yours now!

Download, deploy, and build something we haven't imagined yet. 👀👇
- Hugging Face：
https://t.co/4kaAcqYEVj
- ModelScope：
https://t.co/eRIMZCGkhC

[Leon做了个 @leon_made_this]

@HarriesSteele 牛了，我也来贡献一份力量哈哈哈

[Leon做了个 @leon_made_this]

OpenAI 的 computer history 功能已经出了一天了。我刷了半天，好像也没有看到有比较真实的反馈呢。
怎么回事？大家是都不敢开，还是开了的号已经被带走了呢？👀
不瞒你说，我也在等一个勇士替我踩坑。有没有已经当上赛博小白鼠的朋友出来吱一声？🐭

[Leon做了个 @leon_made_this]

曾几何时，Cursor 也被认为仅是套壳，vscode fork。现在早已被证伪，SpaceX 买的其实是数据飞轮，补齐了从算力到模型再到 grok build/ cursor 分发，再到真实工程数据回流的完整链条。

[Cursor @cursor_ai]

Cursor is now part of @SpaceX.

Today, we have officially closed our acquisition. We will join the @SpaceXAI team to help make Grok the world's most useful AI and improve Grok Build, Grok Bot, Grok API, Cursor, and more.

SpaceX has built some of the most inspiring and impressive technology in the world, and we’re grateful for the opportunity to become part of such a special company. Onwards.

[Leon做了个 @leon_made_this]

本地用户福音了， 3.6-27B就是能单卡跑+接近能干活的质量，如果真的如官方说的能比 3.6 跳大一截，那就绝对本地标杆了。

[Qwen @Alibaba_Qwen]

Less than 2 hours to say hi👋. It's almost time!
See you soon: 👀
https://t.co/rXzUQ4d6WQ

[Leon做了个 @leon_made_this]

Less 'cooking', more like putting out fires in the backend. LMAO.

[Samuel McDonnell @samueljmcd]

Is Anthropic cooking or struggling right now?

Why are they SILENT??

[Leon做了个 @leon_made_this]

@hyclove1 发这种交友贴的频率好像有点高

[Leon做了个 @leon_made_this]

@johnnybonek9r2 你的瓶颈即是我的梦想

[Leon做了个 @leon_made_this]

@ChangFengjx 每年一定要定期做体检的。

[Leon做了个 @leon_made_this]

@zeping312 资金就这么多，不是到这就是到那呗。

[Leon做了个 @leon_made_this]

LOL

[Mia @MiaAI_lab]

What a week for AI looks like

Muse Glimmer 30b
Nemotron 3.5 Lightning
Grok Bot
Grok 4.6
Qwen 3.8-Max weights
GLM-5.3
Qwen3.8-27b

Anything I missed?