IMO the HLE scores without tools seem to align better with real world performance of the models.
To me it feels like the difference between "RL performance" and the pretraining / base "knowledge".
Yes you can RL terminal bench to the moon but does the model hold up on out of distribution tasks?
Kind of like trying to navigate a dark room with a laser light, vs a flashlight. Laser is going to go a lot farther much more efficiently but only if you are already pointing it at the right place.
To me it feels like the difference between "RL performance" and the pretraining / base "knowledge".
Yes you can RL terminal bench to the moon but does the model hold up on out of distribution tasks?
Kind of like trying to navigate a dark room with a laser light, vs a flashlight. Laser is going to go a lot farther much more efficiently but only if you are already pointing it at the right place.