Geek out realtime dashboard release 15!
AFD DeepSeek V4.1 Flash Boost real agentic workflow 2x (draft acceptance grammar bug) Basically you can go to the root of the http url and watch all the nitty gritty performance stats stream in realtime. Actually I found a bug watching
Day 5, relaxed, nothing major happened.
Everything that needs to be released has been released. The remaining Deepseek, GLM, and M3 have a very small chance of coming. What deserves attention is that the System One / Jev route is beginning to gather together: CLM-8B + Ali Decision Model The former is fine-tuned based on qwen3-8B, which is also the favorite Qianwen 🐸 🔥There is also a mysterious model, a cute space bunny model, Space Bunny…
Our internal writing benchmark uses 3 AI judges: Claude, GPT and DeepSeek.
So do they favor their own family? The GPT judge scores OpenAI models 2.3 points higher (out of 100) than the other two judges do. The DeepSeek judge gives DeepSeek models +1.7. The Claude judge gives
I'm getting better and better results out of XiaomiMiMo MiMo-v2.6-flash
Tested with a bit difficult prompt, to create a Koi-Kingdom, difficult cause qwen3.8-flash and deepseek-v4.1-flash did struggle this one. But not MiMo. One-shot, three.js, no external assets, single html
You have reached the end of the archive
All of deepseek