TurboQuant

New LLM compression algorithm by Google

314 votes·5 comments·Mar 25, 2026

About

A set of advanced theoretically grounded quantization algorithms that enable massive compression for large language models and vector search engines.