Multiverse Computing launches Quasar 438B AI model

Multiverse Computing has launched Quasar 438B, a new reasoning artificial intelligence model designed for enterprise agents and large-scale programming tasks. The model is compatible with English and Spanish and has achieved a score of 43 on the Artificial Analysis Intelligence Index, which is noted as the highest score for a European model in that ranking.

Rather than being built from scratch, Quasar 438B utilizes the company’s CompactifAI technology to compress the open-weight GLM-5.2 model from Z.ai. This process reduced the parameter count from 753 billion to 438 billion. While independent testing by Artificial Analysis indicates a drop in intelligence scores—from 53 for GLM-5.2 to 43 for Quasar—the compression provides significant performance benefits, including higher speeds of approximately 183 tokens per second and lower operational costs.

Enrique Lizaso, co-founder and CEO of Multiverse Computing, stated that the model demonstrates that European developers do not have to choose between reasoning capacity and speed. The technology aims to reduce technological dependency by providing high-performance, low-latency options for European organizations.