The new launch offers Unmatched Performance, Silicon Freedom and Seamless access via Compass API.
Core42 has announced the availability of OpenAI’s latest open-weight AI models, including gpt-oss-20B and gpt-oss-120B, on the Core42 AI Cloud platform, instantly accessible through the Core42 Compass API. The deployment enables enterprises, researchers, and developers to run the models on a choice of leading silicon platforms with sovereign, scalable, and high-performance capabilities.
Integrated into Compass API with the flexibility to access a wide spectrum of high-performance compute platforms, Core42 delivers industry-leading inference speeds of up to 3,000 tokens per second per user, enabling real-time AI at global scale while aligning workloads with the optimal infrastructure for price-performance and scalability. This deployment delivers tailored performance for low-latency inference workloads and applications, reinforcing Core42’s commitment to secure and optimized global sovereign-enabled AI infrastructure.
“Core42 AI Cloud, powered by silicon-diverse infrastructure, delivers the flexibility and performance needed for today’s AI workloads. Through the Compass API, organizations can access the latest open-weight AI models and choose the optimal platform to scale transformation, optimize performance and cost, and drive progress across global markets,” said Kiril Evtimov, CEO of Core42 and Group CTO, G42.
Key Benefits of Core42’s Open-Weight Deployment:
• Enterprise-scale performance – Run the fastest, most demanding workloads at global scale, enabling advanced automation, decision-making, and real-time AI experiences.
• Sovereign-ready scalability – Deploy high-performance AI in-country with full sovereign controls, supporting secure operations in regulated sectors such as healthcare, finance, and national security.
• Optimized for committed environments – Deliver fast, scalable AI in-country with sovereign controls for organizations operating under committed infrastructure agreements, ensuring predictable cost and performance.
• Cost-efficient agentic AI – Run agentic workloads at the lowest possible cost while maintaining in-country deployment and sovereign controls, making advanced AI accessible to cost-sensitive use cases.
Available now through Compass API, these models let organizations run and adapt AI locally or in the cloud, with full transparency, fine-tuning, and sovereign deployment options. Customers can align performance, cost, and compliance to their needs.
This release marks a pivotal move toward enterprise AI autonomy. With open-weight access, businesses can shape AI to their unique needs and unlock new possibilities for innovation at scale. This announcement builds on G42’s recent milestones, including the unveiling of a 5GW US-UAE AI campus and the launch of the 1GW Stargate UAE facility as Phase 1 of the project, as well as a $1.5 billion investment from Microsoft in 2024, reinforcing the UAE’s role as a global AI hub.
