Zyphra Inference

Serverless inference service launched with Zyphra Cloud for frontier open-weight language models, optimized for high-throughput, low-latency deployment on heterogeneous accelerator infrastructure.