Multiverse Computing Cuts AI Model Size in Half Without Sacrificing Performance — MULTIVERSE COMPUTING, S.L.

Event summary

Multiverse Computing launched Pulsar 16B, a 16.15B-parameter open reasoning model built on NVIDIA Nemotron architecture.
The model delivers the reasoning performance of leading 30B-class architectures at roughly half the parameter count.
Pulsar 16B outperforms gpt-oss-20B on nearly every measure despite being smaller, with significant improvements in instruction following, function calling, and math reasoning.
The model achieves substantial reductions in memory usage and inference time, making it suitable for lower-memory GPUs or single-node environments.

The big picture

Multiverse Computing's Pulsar 16B represents a significant advancement in AI model efficiency, addressing the long-standing trade-off between model size and performance. This development is particularly relevant for regulated industries where data sovereignty and operational efficiency are critical. The collaboration with NVIDIA underscores the growing importance of hardware-software co-optimization in the AI landscape.

What we're watching

Performance Scaling: How the model's performance will scale with further optimizations and real-world deployment.
Market Adoption: The pace at which enterprises adopt Pulsar 16B for high-concurrency and latency-sensitive applications.
Competitive Dynamics: Whether Multiverse Computing can sustain its lead in model compression technology against larger AI providers.