Post by Bindu Reddy on X
Bindu Reddy@bindureddy
XDeepseek math is an IMO gold medal winning model built on top of their V3.2 base
DeepSeek appears to be testing novel scalable post-training methods and will drop an amazing reasoning model
Open source is going to close the gap to top SOTA models quickly
255 likes25 repliesPosted Nov 28, 2025