Plan to test new models?

[SiyangJ]

Impressive work. With deepseek-r1 and QwQ released, any plans to test these promising competitors to o1?

[betmma]

deepseek-r1:
Conceptual 88.5
Original 89.2
Trap 65.2

[betmma]

o3-mini-medium:
Conceptual 93.1
Original 92.5
Trap 51.0
o3-mini-high:
Conceptual 93.8
Original 91.7
Trap 50.3
(note that it's tested on o3-mini api, and it doesn't return reasoning tokens)