# Code-Optimized Tokenizer

A large multilingual BERT model developed by the Algomatic team, supporting 8192 context length, trained on approximately 60 billion tokens, suitable for mask filling tasks.