Spanish Firm Multiverse Computing Releases Quasar 438B AI Model
Spanish AI company Multiverse Computing has launched Quasar 438B, a 438-billion parameter reasoning model featuring a 1-million token context window. It scored 43 on the Artificial Analysis Intelligence Index v4.1.1, making it the top-ranking European model on the benchmark. This release reflects Europe's expanding enterprise AI capabilities, surpassing regional competitors like Mistral Medium 3.5 on key intelligence metrics. However, notes suggesting the model may be built upon Zhipu AI's GLM5.2 underscore the ongoing debate around European AI autonomy and base model provenance. Available via the CompactifAI API at $0.60 per million input tokens and $1.80 per million output tokens, Quasar 438B is a text-only proprietary model that generates output at 176.2 tokens per second. It also scored 75.0 in long-context reasoning and 69.3 on Terminal-Bench v2.1 for agentic terminal tasks.
## BACKGROUND
Artificial Analysis provides independent benchmark evaluations for artificial intelligence models, combining tests across math, coding, and reasoning into a aggregate score. Multiverse Computing is a Spanish technology firm known for quantum-inspired algorithms and model compression tools designed to make large language models more efficient.