I had trust issues with all the Jev demos so I had to see for myself if the performance gains we're real.
Here I have a 128 person tennis bracket and ran Deepseek v4 against Jev to compare the results. Given the right use case Jev is game changer 🧵
GLM 5.3 Flash
Vs. MiMo V2.6 Flash vs. DeepSeek V4.1 Flash @ 2x DGX Sparks The baseline for Flash models has raised so much across the board.
Switched to DeepSeek v3 Pro.
Solid architecture and token optimization that puts it on top. It's still taking time in the new version, they haven't resolved that yet.
You read that right, the open source model also has the Pirate Bay🤩, and someone made the open source model on Hugging Face a seed🧲
I clicked on it and looked at DeepSeek-V4.1-Flash: 765GB, 88 files, Hugging Face’s official SHA-256 hanging next to each weight file; There is a Magnet in the upper right corner, and 20 people are currently planting seeds. You can easily register your own handler and bind it to HF…
You have reached the end of the archive
All of deepseek