Amazon saves millions by replacing Sonnet 4.6 with Strands Decider

Amazon replaced Sonnet 4.6 with Strands Decider, a decision model, to label tickets in their internal system, resulting in a cost savings of 98% at the same level of performance. The new model, E4B Gemma 4, was chosen after testing and tuning, and was deployed on SageMaker using ml.g5.xlarge instances. The model uses 19 GiB of memory and takes around 640 ms to process a request. The change is expected to save millions of dollars in token costs annually.