I'm building an outbound caller when latency is a key consideration. 4o mini handles the conversation well but doesn't consistently trigger tools as needed. Typically, I'd opt for 4o, but it's latency is just a little too high for this use case. Has anyone has success with the 4.1 models, 4.1 mini is an improvement tools wise, but I'm cautious given it's advertised latency (even though It doesn't seem to respond slower then 4o while testing). Overall, any model recommendations to try out would be greatly appreciated as my LLM knowledge is limited.