The number I am watching is not a trillion and a half. It is three letters: MIT.
That is the licence that says do whatever you like, just leave our name somewhere. No user threshold, no ban on commercial use, no clause about competitors.
Fact
The facts, per DeepSeek's official profile on Hugging Face: DeepSeek-V4-Pro-0813 was published on 13 August 2026, with 1650 billion parameters and an MIT licence. The weights are in FP8 format, that is eight bits per number instead of sixteen. The model is for text generation and conversation.
Who can actually use it
Not you, if you were thinking of standing it up on your own machine. One thousand six hundred and fifty billion parameters want a hall, even at eight bits per number.
The licence, though, is written for the middlemen. Any compute provider can stand it up and sell it by the hour, asking nobody and sharing no revenue. From there the model reaches you at a price somebody else worked out.
A permissive licence does not open the model to the person at home. It opens the market of the people reselling it.
And that is exactly the move. Hand out weights with a licence that gets in nobody's way, and a dozen providers will have it up within a week, so your model is suddenly everywhere. The closed competitor has to build the same thing alone.
The eight bits
FP8 means every number takes half the space. The files are smaller, the memory stretches further, the speed goes up. You pay in a little accuracy, which at this size is barely felt.
So they are shipping it in the form that is cheapest to stand up. That is not an accidental decision; it is part of the same calculation.
For you today: look at which provider it comes through and at what price. With a licence like this there is usually more than one option, and the gap between them is not small.
Official primary sources
DeepSeek-V4-Pro-0813, DeepSeek's official profile (Hugging Face)The visual is generated code-art, no third-party images.