A startup claims it broke through a bottleneck that’s holding back LLMs
Miami-based startup Subquadratic has announced a new large language model (LLM) called SubQ, claiming it significantly outperforms others in speed, cost, and energy efficiency. The company has shared independent test results that suggest its model can process up to 12 times as much text as competing models while matching their performance on key tasks. While reactions include skepticism about its claims, further verification of SubQ could change perceptions in the AI community.
This may affect professional readers if the source supports the claimed AI impact.
This record is extracted from a published AI Today issue and tied to the original source URL. Treat the source as the record of evidence for the summary.