Comment by hintymad

1 day ago

My guess is that Gemini team didn't focus on the large-scale RL training for the agentic workload. And they are trying to catch up with 3.1.

0 comments