Hacker News new | ask | show | jobs
by kirtivr 23 days ago
They got 1K tok/s with Deepseek v4 Pro. That's kinda cool..
3 comments

No they didn't, they predict they'll get that much. Also worth noting the prediction assumes running at MXFP4/FP8 quantization.
Exactly! Any optimization for local inference is a welcome change IMHO!
Thanks. To be fair, this number is what we expect to get once we port DeepSeek V4 in our engine on the upcoming generation of GPUs!