Multiverse Computing Launches Pulsar 16B in collaboration with NVIDIA: Frontier-Grade Reasoning at Half the Parameters
AI Platform, Partnership, Product
Positive
Multiverse Computing has launched Pulsar 16B, an open reasoning model built on NVIDIA's Nemotron architecture. The model was developed using NVIDIA Model Optimizer and Megatron infrastructure, and was independently validated on NVIDIA accelerated computing hardware.
The release highlights the continued adoption of NVIDIA's developer tools and accelerated computing platform by third-party AI model builders. The collaboration positions NVIDIA's infrastructure as a key enabler for frontier-class AI reasoning models targeting efficiency-focused deployment scenarios.
Why it matters
The collaboration demonstrates demand for NVIDIA's model development ecosystem — including Nemotron architecture and Model Optimizer tooling — reinforcing the company's position as a foundational platform for third-party AI development. Broad adoption of NVIDIA tools by independent AI firms supports the durability of its enterprise software and hardware ecosystem.
Key facts
Pulsar 16B is built on NVIDIA's Nemotron architecture • Development utilized NVIDIA Model Optimizer and Megatron infrastructure • The model was independently validated on NVIDIA accelerated computing hardware • Multiverse Computing is headquartered in Donostia, Spain