We introduce the Byte Latent Transformer (BLT), a new byte-level LLM architecture that, for the first time, matches tokenization-based LLM performance at scale with significant improvements in inference efficiency and robustness. BLT encodes bytes into dynamically sized patches, which serve as the primary units of computation. Patches are segmented dynamically based on the entropy of the next byte, allocating more compute and model capacity where increased data complexity demands it. We present the first flop controlled scaling study of byte-level models up to 8B parameters with 4T training bytes. Our results demonstrate the feasibility of scaling models trained on raw bytes without a fixed-vocabulary. Both training and inference efficiency improve due to dynamically selecting long patches when data is predictable, along with qualitative improvements on reasoning and long tail generalization. Overall, for fixed inference costs, BLT shows significantly better scaling than tokenization-based models, by simultaneously growing both patch and model size.
AUTHORS
Written by
Artidoro Pagnoni
Ram Pasunuru
Benjamin Muller
Margaret Li
Chunting Zhou
Lili Yu
Gargi Ghosh
Mike Lewis
Ari Holtzman
Srini Iyer