The cost of a Benchmarking job depends on the size of the sample dataset used for evaluation, the number of machine translation services that are compared, and the benchmark metrics that are calculated. 

For instance, if you would like to benchmark three MT services with varying cost of 10-20$ per million character, you would be required to spend 500$ for a ten-million character dataset that is used for benchmarking. 

Those parameters you can be fully configured based on the number of models, the size of the dataset, and the performance metrics that you are benchmarking using. You will always receive an estimate of the total benchmarking cost while selecting those configurations.