Google DeepMind researchers published research on test-time compute in August and proposed this technique as a potential way to keep large language models improving through the peak-data wall.
Results that may be inaccessible to you are currently showing.