The AI Large Model Intellectual Disability Competition is here😑 The contestants this time: - gpt-6-luna
Gpt-5.6-luna -grok-4.6 - deepseek-v4-pro - kimi-k3 -gemini-3.1-pro The following are all based on the highest thinking level test of this model. Choose the one you think is the most retarded! !
DeepSeek Harness 桌面版的安装包被网友扒出来了 🤣 官方还没发公告，很多人已经吃上螃蟹了（包括我），我这有最新的 0.1.7-rc.2 下载地址~
优点是对话记录和配置都和网页版同步，而且比较意外的是桌面端和网页版的 UI 和插件有点儿区别，还可以登录 DeepSeek 账号查看余额。
Another three.js XiaomiMiMo MiMo-v2.6-flash example.
This is way closer to deepseek-v4.1-flash result in terms of world, but overall graphics is better here. Three.js, one-shot, direct api call with thinking high. I'm loving v2.6-flash model results so far.
There's something about this! A personal AI investment research workbench that turns A-share/U.S./Hong…
There's something about this! A personal AI investment research workbench that turns A-share/U.S./Hong Kong stock investment research into evidence-based Vibe-Research, 2.5K stars, 117 data endpoints + six-stage research SOP, leaving a chain of evidence at every step, no longer trusting the black box that the model just opens its mouth. In the past, when AI was used for investment research, it would either give you…
You have reached the end of the archive
All of deepseek