If you are waiting for tons of tool calls before seeing anything happen, just know that agent behaves much differently with Sonnet (I’m using 3.7).
Talking about agent behavior, not base model performance:
Sonnet: 0-2 min
4o-mini, o3: 2-20 min
If you are waiting for tons of tool calls before seeing anything happen, just know that agent behaves much differently with Sonnet (I’m using 3.7).
Talking about agent behavior, not base model performance:
Sonnet: 0-2 min
4o-mini, o3: 2-20 min