@JingyuTang621
First-year PhD student at UT Austin
AI
A creator says the best model reliably completes only 36% of the benchmark’s tasks.
AI
X
