I tested DeepSeek V4 Flash with Vision (DS4FV) using the two DGX Sparks together with TP‑2, with four staggered requests.
It took 57.14 s to complete all four, compared to 36.35 s for Q38FN and 83.87 s for GLM53F. If you mean two independent instances of DeepSeek, one per
Yesterday, DeepSeek announced a temporary trial version of DeepSeek V4.1 Flash in the WeChat group.
Now we can add a model configuration to DeepSeek Harness and use it. I tried it myself and the speed is very fast, reaching 300-400 tps. In the past, when I was thinking about it in max mode, there was a lot of crackling output, but now even if I turn on max
I recently used an AI assistant, I named him Xiaofei.
A few days ago, I linked my X account to him that I have used for more than five years. It didn’t take long for him to sort out the AI Coding, video generation and speech recognition that I had been paying attention to for a long time from the content, likes and bookmarks I had posted.
Found a FREE AI coding agent you should try.
Alysis Code can plan, build, debug, test and review your code directly from your terminal. And the interesting part: it comes with free credits, is open-source, and lets you bring your own AI model. Supported models/providers
You have reached the end of the archive
All of deepseek