Multiverse Computing Launches Quasar 438B
aiinsider ·
Multiverse Computing has launched Quasar 438B, a 438-billion-parameter bilingual reasoning model designed for enterprise agents and coding. The model supports English and Spanish and scored 43 on the Artificial Analysis Intelligence Index v4.1.1, the highest score among European models in the comparison. It generated 500 output tokens in 15.3 seconds, outperforming Mistral Medium 3.5 and NVIDIA Nemotron 3 Ultra on the reported score and speed measures. Quasar scored 75.0 on the Artificial Analysis Long Context Reasoning evaluation and 69.3 on Terminal-Bench v2.1, demonstrating capabilities in long-document analysis, coding, and command-line tasks. The model is available through the CompactifAI API for applications including software engineering, operational automation, document-heavy research, and technical knowledge work. Multiverse Computing said it plans to continue developing Quasar’s coding and agentic capabilities.
aiinsider