Her 7-million-parameter model achieves 45% on ARC-AGI-1.
Alexia was the solo author on TRM, pushing the limits with recursion on a small 2 layer network
TRM’s performance on ARC-AGI is incredibly efficient
Built on tops of HRM from @makingAGI
Congratulations on the new role @jm_alexia
I'm excited to share that I joined Microsoft as Principal researcher. Microsoft #1! 🚀
No Digg Deeper questions have been answered for this story yet.
@jm_alexia Congratulations, Alexia! Microsoft is lucky to have you.
I'm excited to share that I joined Microsoft as Principal researcher. Microsoft #1! 🚀
https://arxiv.org/abs/2510.04871
https://arxiv.org/abs/2506.21734
https://m.youtube.com/watch?v=P9zzUM0PrBM&source_ve_path=MjM4NTE&embeds_referring_euri=https%3A%2F%2Farcprize.org%2F
Alexia was the solo author on TRM, pushing the limits with recursion on a small 2 layer network
TRM’s performance on ARC-AGI is incredibly efficient
Built on tops of HRM from @makingAGI
Congratulations on the new role @jm_alexia
