CertBrain
Courses
Home
Blog
Sign in
← Tokens, Chunking, Embeddings and Vectors
1 / 11
Question
A developer notices that the same 500-word paragraph counts as a different number of tokens on two different models in Amazon Bedrock. What explains this?
1
Token counts depend on which AWS Region the model is running in today
2
Each model uses its own tokenizer, which splits text differently
3
Token counts change each time the same text is sent to a model
4
One of the models is miscounting and should be reported to AWS
Press 1–4 to pick an answer
CertBrain