AI Hub
All models
LLMOpen weights

BERT

Google

Updated May 21, 2026 · Benchmarks via Artificial Analysis, specs & pricing via OpenRouter · Methodology · Report an error

Specifications

Type
LLM
Access
Open weights
Released
October 11, 2018
Parameters
340M
Context window
512 tokens
Input
Text
Output
Text

Our take

A bidirectional Transformer encoder that set new records across NLP and made pretrain-then-fine-tune the universal recipe. For years it was the most influential model in the field.

Related research

Compare BERT with

See all BERT alternatives →