Kimi K3.1 Pro is the smartest Chinese model.
It's also the slowest one on this chart Best model from 7 Chinese labs: Kimi K3.1 Pro: Index 46.5, 28 tokens/s DeepSeek V4.1 Flash: 32.6, 157 tokens/s Step 5 Preview: 28.5, 134 tokens/s Only DeepSeek scores above 30 AND tops
20 days ago we built a system that writes GPU code.
Today its kernels outperform the state of the art. We just open-sourced 9 megakernels, which outperform NVIDIA’s stack on DeepSeek and run self-driving, robotics, protein folding and world models 2–3× faster than the current
Catalog note, Built in Venice: the same snapshot picked up DeepSeek V4.1 Flash 1017 Abliterated - E2EE, private, 1M context.
Uncensored and private on the same route.
Two private models, now included with Pro.
DeepSeek V4.1 Flash and GLM 5.3 Flash run in Private mode, so no prompt or response is stored. Try it on Venice.
You have reached the end of the archive
All of deepseek