AWS Certified AI Practitioner AIF-C01 Practice Question

Because the application must respond almost instantly, the most important consideration is each model's expected inference latency. Latency directly influences how quickly the model can generate a response; if latency is too high, the assistant will feel sluggish to users. While model size, multi-lingual support, and maximum input/output length are relevant in other situations, they do not address the startup's primary need for sub-second responsiveness.