Quasar 438B: Europe's Top AI Model, Outperforms Rivals

Multiverse Computing has launched Quasar 438B, a new flagship reasoning model. This model achieves the highest score for a European model on the Artificial Analysis Intelligence Index, reaching 43. It also boasts a 15.3-second response time for enterprise agents and coding tasks. Quasar is the first large model from Multiverse Computing, supporting both English and Spanish. It outperforms Mistral Medium 3.5, which scores 30, and NVIDIA Nemotron 3 Ultra, which scores 38. The model produces 500 output tokens in 15.3 seconds, making it faster than Mistral Medium 3.5, which takes 18.8 seconds. What's interesting is that Quasar also scores 75.0 on Artificial Analysis Long Context Reasoning, matching Grok 4.6. On Terminal-Bench v2.1, it scores 69.3, ahead of Mistral Medium 3.5 and Nemotron 3 Ultra. This shows its ability to handle complex coding and command-line tasks. The bottom line is this new model strengthens Europe’s sovereign AI capability, demonstrating competitive performance and speed.