
Allen Institute for AI Unveils Bolmo 7B and 1B, Pioneering Open Byte-Level Language Models
The Allen Institute for AI has introduced Bolmo 7B and Bolmo 1B, touted as the first fully open byte-level language models. Built on the Olmo 3 architecture, these models eliminate the need for tokenizers, enhancing multilingual processing capabilities and efficiency in training. They are designed to perform competitively in noisy or low-resource environments, marking a significant advancement in language model technology.


